Direction des Études et Synthèses Économiques
G 2014 / 05
The Impact of Hazardous Industrial Facilities on
Housing Prices: A Comparison of Parametric and
Semiparametric Hedonic Price Models
Céline GRISLAIN-LETRÉMY et Arthur KATOSSKY
Document de travail
Institut National de la Statistique et des Études Économiques
INSTITUT NATIONAL
DE LA
STATISTIQUE
ET DES
ÉTUDES ÉCONOMIQUES
Série des documents de travail
de la Direction des Études et Synthèses Économiques
G 2014 / 05
The Impact of Hazardous Industrial Facilities on
Housing Prices: A Comparison of Parametric and
Semiparametric Hedonic Price Models
Céline GRISLAIN-LETRÉMY* et Arthur KATOSSKY
MAI 2014
Les auteurs remercient Pierre-Philippe Combes, Éric Dubois, Laurent Gobillon, Nicolas
Grislain, Anne Lafferère, Claire Lelarge, David Martimort, Philippe Mongin, Corinne Prost,
Sandrine Spaeter, Corentin Trevien et Bertrand Villeneuve pour leurs remarques et
suggestions. Ce travail a également bénéficié de commentaires de Pauline Charnoz, Amélie
Mauroux et des participants au 28ème congrès de l’association économique européenne
(EEA), au 12ème atelier international de la statistique et de l’économétrie spatiales, au
séminaire de recherche de l’Insee « Immobilier en France : analyses et perspectives », aux
conférences de 2012 de l’association européenne de l’économie des ressources et de
l’environnement (EAERE) et de l’association de science régionale européenne (ERSA) et
aux 11èmes journées Louis-André Gérard Varet. Les auteurs remercient la Direction
générale de la Prévention des risques, notamment Sandrine Robert et Grégory Dubois, ainsi
que la Direction générale de l’Aménagement, du logement et de la nature du ministère de
l’Écologie pour l’obtention des données. Ils remercient tout particulièrement Vincent Binet,
Rémi Borel, Olivier Dupret, François Filior, Martine Giloppe, Jeanne-Marie Gouiffès et
Brigitte Pouget des centres d’études techniques de l’équipement (CETE) Normandie-Centre,
Nord-Picardie et Sud-Ouest pour leur aide dans le recueil des données. Les auteurs
remercient enfin Christophe Yon pour avoir réalisé les trois cartes fournies dans ce
document de travail.
_____________________________________________
*
Département des Études Économiques - Division « Marchés et entreprises » Timbre G230 - 15, bd Gabriel Péri - BP 100 92244 MALAKOFF CEDEX
Département des Études Économiques - Timbre G201 - 15, bd Gabriel Péri - BP 100 - 92244 MALAKOFF CEDEX - France - Tél. : 33 (1) 41 17
60 68 - Fax : 33 (1) 41 17 60 45 - CEDEX - E-mail : d3e-dg@insee.fr - Site Web Insee : http://www.insee.fr
Ces documents de travail ne reflètent pas la position de l’Insee et n'engagent que leurs auteurs.
Working papers do not reflect the position of INSEE but only their author's views.
The Impact of Hazardous Industrial Facilities on Housing
Prices: A Comparison of Parametric and Semiparametric
Hedonic Price Models
Abstract
Households’ willingness to pay for prevention against industrial risks can be revealed by real
estate markets. With highly detailed microdata, we study housing prices in the vicinity of
hazardous industries located near three important French cities (Bordeaux, Dunkirk and
Rouen). We show that the impact of hazardous plants on housing values strongly differs
among the three studied areas, even if they all surround chemical and petrochemical
industries. We compare the results from standard parametric hedonic property models and
from a more flexible, semiparametric hedonic property model. We show that the parametric
model may structurally lead to an important bias in the estimated value of the impact of
hazardous plants on housing values and in its variations with respect to distance to the
plants.
Keywords: hedonic analysis, locally-weighted regression, urban housing markets,
industrial risk
L'impact des industries à risques sur le prix des
logements : comparaison de modèles paramétriques et
semi-paramétriques de prix hédoniques
Résumé
Le consentement à payer des populations pour des mesures de prévention contre les
risques industriels peut être révélé par les marchés immobiliers. Avec une base de données
microéconomiques très détaillées, nous étudions les prix des logements à proximité
d’industries dangereuses situées près de trois grandes villes françaises (Bordeaux,
Dunkerque et Rouen). Nous montrons que l’impact des usines dangereuses sur le prix des
logements varie beaucoup entre les trois zones étudiées, même si elles sont toutes à
proximité d’industries chimiques et pétrochimiques. Nous comparons les résultats d’un
modèle paramétrique standard de prix hédoniques et ceux d’un modèle de prix hédoniques
semi-paramétrique, plus flexible. Nous montrons que le modèle paramétrique peut
structurellement conduire à un biais important dans la valeur estimée de l’impact des usines
dangereuses sur le prix des logements et dans ses variations en fonction de la distance aux
usines.
Mots-clés : analyse hédonique, régression localement pondérée, marchés du logement
urbains, risque industriel
Classification JEL : C21, Q51, R52, R21
2
1
Introduction
Real estate markets can reveal households’ willingness to pay to reduce their exposure
to hazardous industrial facilities. Clearly, hazardous industrial activities generate strong
externalities, many of them being negative. Indeed, neighboring populations partly bear
the cost of a potential industrial accident,1 and may endure day-to-day nuisances associated with the ordinary course of activity. However, these facilities also generate positive
externalities (Greenstone et al., 2010), potentially more diffuse: they provide, directly and
indirectly, employment; through local taxes, they may contribute to the economic development of the municipalities. Declining the impact of distance to hazardous facilities on
housing values with respect to their activities reveals in which extent these activities are
perceived as disamenities. We estimate the impact of distance to hazardous plants on
housing values by using hedonic price models. The first-order derivative of the hedonic
price function with respect to the distance to hazardous plants provides an estimate of the
implicit price of the distance to these plants, that is an estimate of the buyers’ willingness to pay to live far from these plants. We study housing prices in the vicinity of three
important French cities (Bordeaux, Dunkirk and Rouen). These three chosen industrial
areas all include chemical and petrochemical industries, but with different socioeconomic
characteristics of their neighborhoods and different perceptions of industrial risks.2
Our first contribution is to study the impact of hazardous chemical and petrochemical
industries on housing prices with highly detailed microdata. Our unique database is much
richer than the ones used for studies relative to the impact of similar industrial risks on
housing prices. Impact of risk exposure on real estate prices is estimated for oil facilities
by Boxall et al. (2005) and Flower and Ragas (1994), for chemical industry by Carroll
et al. (1996), and for industrial areas with chemical or petrochemical facilities by Sauvage
(1997) and Travers et al. (2009). These studies use data with few (or without) extrinsic characteristics of the dwellings and without information on buyers and sellers.3 Our
1
Damages certainly imply the liability of the industrialist, but compensation can be delayed and remain
partial. In particular, some physical and moral damages cannot be repaired.
2
The industrial areas near Bordeaux and Rouen have become sadly well-known after our data collection.
In the gunpowder near Bordeaux, a fire caused one death and two seriously injured persons in December
2013 (“Gironde : un mort et deux blessés dans un incendie sur un site classé Seveso” by Jean-Pierre
Tamisier, December 6th, 2013, Sud-Ouest). Near Rouen, mercaptan gas escape from one of the factories,
the Lubrizol company, caused foul odor in France and in England in January 2013 (“Foul-Smelling Cloud
Drifts Over France, Alarming Residents” by Scott Sayare, January 22nd, 2013, The New York Times).
3
See Table 13 in Appendix A.2 for a review of all these studies and their data. The impact on housing
prices of other industrial activities has been studied, in particular the impact of power plants (Davis,
2011), waste treatment and storage (Greenstone and Gallagher (2008), see Farber (1998) for a review) and
of natural gas (Boxall et al., 2005). Air pollution also significantly decreases housing prices (Chay and
3
case study relies on a detailed dwelling-level micro-database, each dwelling’s address being precisely geocoded. Detailed data relative to their extrinsic characteristics have been
collected: proximity to central business district, shops and public utilities, exposure to
industrial risk, to other risks or pollutions. Dwellings’ price and intrinsic characteristics
come from notarial data; information about buyers and sellers is also provided.
Our second contribution consists in comparing the impacts of distance to hazardous plants
on housing prices, as estimated by two models: a standard parametric model and a semiparametric model. The semiparametric regression is a locally weighted regression that
allows implicit prices to vary with respect to space, time of sale and buyer’s characteristics, while keeping some smoothness in their distribution.4 It enables to relax and to test
the assumption of fixed parameters, that is of constant implicit prices across space, time
and among buyers. The fixed-parameter assumption is rejected, which confirms the need
for flexible forms and pleads for the use of semiparametric models. Such a comparison of
parametric and semiparametric models had not been performed until now in the analysis
of industrial risks, but for other amenities, such as agricultural pollution on housing prices
(Bontemps et al., 2008), light rail access (Redfearn, 2009); all these analyses plead for the
use of semiparametric models. Anglin and Gencay (1996) also show that semiparametric
hedonic models outperform parametric ones. McMillen (2010), McMillen and Redfearn
(2010), Redfearn (2009), Sunding and Swoboda (2010) specifically compare parametric
models with locally weighted regression and recommend the use of this semiparametric
model. Redfearn (2009) shows that implicit prices vary spatially and temporally and
that assuming fixed implicit prices is a misspecification of hedonic model. This misuse of
parametric models has important consequences in our case study. If signs and orders of
magnitude of effects are similar in the two models for the very wide majority of coefficients,
the estimated impacts of the distance to highly hazardous plants on housing prices significantly differ between the two models. The parametric model leads to an important bias
(here an overestimation) in the estimated value of this impact near Bordeaux and Rouen
and in its variations with respect to distance to the plants near Rouen.
Using the semiparametric model, we show that the impact of hazardous plants on housing
values strongly differs among the three studied industrial areas, even if they all surround
Greenstone (2005), see Maslianskaïa-Pautrel (2008) for a review).
4
Locally weighted regressions have already been applied in several works by the French National Institute of Statistics and Economics Studies (INSEE). See Floch (2012) for examples and a methodological
presentation of this model in French.
4
chemical and petrochemical industries. The gunpowder factory near Bordeaux is a former
military plant, not necessarily perceived as hazardous by neighboring population. We find
that its proximity is even valued in its very close vicinity, probably because its neighborhoods are green and very quiet places. Thus, we capture here unobserved amenities which
are spatially correlated with the distance to the plant. Near Dunkirk and Rouen, chemical
activities are clearly identified as hazardous by local populations. However, near Dunkirk,
we find no significant impact of the distance to highly hazardous plants on housing prices,
likely because these industrial risks are overshadowed by the nuclear plant in Gravelines
(located 18 km from Dunkirk). Near Rouen, highly hazardous plants are perceived as
disamenities: on average, households are willing to pay around 1.2% of their dwelling price
to go 100 more meters away from these plants. This marginal willingness to pay decreases
with respect to the distance to the plants, as marginal gains (in terms of exposure reduction) of going further away from the hazard source are likely to decrease. Finally, we
show that this marginal willingness to pay increases over time following the AZF accident,
information policies, or the implementation of the technological disasters insurance system.
Our findings have important implications. Inasmuch as households are aware of the differences in amenities across locations,5 the impact of distance to highly hazardous plants on
housing prices can be interpreted as the local populations’ marginal willingness to pay for
prevention against industrial risks. This willingness is used in cost-benefit analyses that
estimate the efficient level of prevention against industrial risks for the civil society (see
Treich (2005) for a review). The benefits, which correspond to the population’s willingness
to pay for this reduction, are often revealed by real estate markets.6 Most of the time
this willingness to pay is estimated by parametric hedonic models and for similar types
of industrial activities, over a potentially different study area or period. Our results show
that parametric hedonic models can lead to an important bias in the estimated value of
the marginal willingness to pay. Besides, our results show that estimated willingness to
pay for prevention has a limited external validity: it strongly differs among industrial areas, even among chemical and petrochemical industries; it also depends on the distance
5
Currie et al. (2013)’s findings do not contradict perfect, or at least unbiased, information about industrial hazardous activities in the housing market, even in the presence of scientific uncertainty about health
risks.
6
The benefits are the monetary value of the reduction of risk exposure, that is the population’s willingness to pay for this reduction. This willingness to pay can be assessed either by stated or revealed
preference methods. The stated preference method comes up against numerous biases in the questionnaire
of so called contingent valuation studies and the limited incentives of questioned households to reveal their
preferences. The revealed preference method consists in observing actual individual decisions on markets,
such as real estate market.
5
to these facilities and on time. Thus, our findings plead for a careful use of population’s
willingness to pay for prevention against industrial risks in the cost-benefit analyses of
prevention measures, as it can lead to a significant bias in the estimation of the efficient
prevention level.
The paper is organized as follows. The parametric and semiparametric models are presented in Section 2. Section 3 describes the three industrial areas, the delimitation of study
areas and the data. Section 4 exposes and compares the results get from parametric and
semiparametric models. Section 5 concludes.
2
Model
Hedonic property models. Real estate markets can reveal households’ willingness to
pay to reduce their exposure to hazardous industrial facilities. Hedonic property models
enable to estimate the implicit price of the distance to highly hazardous plants (which is
the first-order derivative of price with respect to the distance to highly hazardous plants).
This implicit price is equal to the households’ (buyers’ or sellers’) marginal willingness
to go one more meter away from these plants. Indeed, in the framework formalized by
Rosen (1974), a dwelling is defined by its distance d to highly hazardous plants and several
other characteristics X, which determine its price P (d, X). When choosing their location,
households equalize their marginal willingness to pay for increasing each characteristic by
one unit with the marginal, or implicit, price of this characteristic.7 Thus, estimation of
the hedonic price function provides an estimation of the implicit price of the distance d to
highly hazardous plants ∂P (d, X)/∂d, which can so be interpreted as households’ marginal
willingness to pay to go one more meter away from highly hazardous plants.
Unless making very specific assumptions, the hedonic price function is not linear and has
no known explicit form (see Freeman (2003) for a review of hedonic price methods). This is
the reason why we have first performed many parametric models: linear, log-linear, log-log
and linear with Box-Cox transformations of the price and continuous regressors - while
7
Formally, a household of income y chooses his location by maximizing his utility U (z, d, X), where z
denotes the amount of composite consumer good (which includes all consumer goods except land), under
his budget constraint y = z + P (d, X). Location choices by households maximize their utility by equalizing
their marginal rate of substitution between each characteristic (d or xk ) and money with the implicit price
(∂P (d, X)/∂d or ∂P (d, X)/∂xk ):
Ud (z, d, X)
∂P (d, X)
=
,
Uz (z, d, X)
∂d
∀k,
6
Uxk (z, d, X)
∂P (d, X)
=
.
Uz (z, d, X)
∂xk
(1)
allowing these Box-Cox coefficients to be different (Grislain-Letrémy and Katossky, 2013).
For each of these specifications, we have either added fixed effects (municipality dummies)
and tested for different types of spatial dependency (spatial errors and/or spatial lag).
Comparison by likelihood ratio tests of these nested models shows that a more flexible
form is indeed better (Grislain-Letrémy and Katossky, 2013).
In addition to the baseline log-linear model, we present here two extensions of this model:
a parametric model and a semiparametric one.
2.1
Parametric model
The parametric model is a log-polynomial model with municipality and time dummies
based on ordinary least squares. We estimate the logarithm of the price Pi of dwelling i
as a function of the distance to the plants di , the dwelling’s other characteristics Xi and
the vector ηti of time dummies (for month and year of sale). Near Bordeaux, the square
and the cube of the distance to highly hazardous plants are significant and included in
the regression. Near Rouen, only the square of the distance to highly hazardous plants is
significant and included in the regression.
OLS Bordeaux: ln Pi = α + β1 di + β2 d2i + β3 d3i + γ 0 Xi + τ 0 ηti ,
OLS Dunkirk: ln Pi = α + β1 di + γ 0 Xi + τ 0 ηti ,
OLS Rouen: ln Pi = α + β1 di + β2 d2i + γ 0 Xi + τ 0 ηti .
2.2
(2)
(3)
(4)
Semiparametric model
Motivation. The semiparametric model is also based on a log-linear specification. It is
a locally weighted regression (LWR), which allows the marginal willingnesses to pay for
characteristics to vary with respect to space, time of sale and buyer’s characteristics, while
keeping some smoothness in their distribution. Indeed, the marginal willingness to pay for
each characteristic is a priori not uniform over the study area (McMillen, 2010). It may
also vary after events that could have changed risk perception, such as local or national
accidents, information policies. Indeed, several empirical works show that real estate prices
can be significantly modified by information policies,8 or insurance coverage.9 Finally, this
8
See Gayer et al. (2000) and Kohlhase (1991) in the case of hazardous waste facilities and Maani (1991)
in the case of high-pressure gas pipeline. However, in the case of airport noise disclosure, Pope (2008)
shows that publicly available information may not be adequately considered by all buyers.
9
For example, flood insurance shapes real estate prices (MacDonald et al. (1990), Harrison et al. (2001),
Morgan (2007) and Bin et al. (2008)).
7
willingness to pay may vary among buyers, because of heterogenous preferences.10
McMillen (2010), McMillen and Redfearn (2010), Redfearn (2009), Sunding and Swoboda
(2010) specifically compare parametric models with locally weighted regression and plead
for the use of this semiparametric model. Redfearn (2009) provides two main reasons that
plead for the use of locally weighted regression instead of standard hedonic models: the
rejected standard assumption of fixed implicit prices and the presence of omitted local
amenities. First, Redfearn (2009)’s results show that the standard assumption of fixed
implicit prices is rejected and suggest that imposing fixed parameters generates spatial
patterns in the errors that leads to parameter estimates highly sensitive to very small
changes in sample or in specification. Second, Redfearn (2009) compares regressions with
and without a known local amenity and shows that the local regression analysis appears
more robust to omitted local amenities.
Locally weighted regression: formalization. We estimate the logarithm of the price
Pi of dwelling i as a function of the distance to the plants di , the dwelling’s other characteristics Xi and the vector ηti of dummies for time of sale. We allow the coefficients
to vary with respect to space (x and y coordinates, xci and yci ), time of sale (ti ) and
buyer’s income (yi ), while keeping some smoothness in their distribution. We choose here
the income as the main buyer’s characteristic. Indeed, income is imputed using gender,
age, marital status and municipality of origin (Section 3) and it summarizes this way the
main observed buyer’s characteristics.11 By denoting Zi = (xci , yci , ti , yi ), we get
LWR: ln Pi = α(Zi ) + β(Zi )di + γ 0 (Zi )Xi + τ 0 (Zi )ηti .
(5)
More precisely, locally weighted regression is a set of weighted least square regressions,
with one regression for each observation (see McMillen and Redfearn (2010) for a review).
Each regression estimates the implicit prices at each observation using a subsample of
“close” observations. Proximity refers here to spatial proximity, temporal proximity, as
well as buyer’s characteristics proximity. The set of observations used in each local regression corresponds to the observations within a window around the considered observation
10
Proximity in terms of preferences may be already partly captured by geographical proximity, because
of a sorting on housing market.
11
We could have also allowed the marginal willingnesses to pay for characteristics to vary with respect to
other characteristics of buyers, but the results get from parametric models provide very little hope in this
direction. Indeed, in several parametric models we did not get significant coefficients when the distance to
the plant was crossed with other characteristics of buyers.
8
j.12 Formally, the objective function for estimation of locally weighted regression is, for
observation j:
LWR:
n
X
(ln Pi − α − βdi − γ 0 Xi − τ 0 ηti )2 Wij ,
(6)
i=1
where Wij = f (Zi , Zj ) is the weight for each observation i; this weight is a decreasing
function of the spatial, temporal and buyer’s distance from observation i to the considered
observation j.
Weights. First, we apply the Mahalanobis distance dM (Zi , Zj ) between Zi and Zj to go
from four dimensions to one dimension in our kernel:13
q
d (Zi , Zj ) = (Zi − Zj )T (cov(Z))−1 (Zi − Zj ),
M
(7)
where Z = (Z1 , ..., Zn ). Then, weights are generated by a kernel weighting function, here
the tri-cubic kernel weighting function, applied to this distance:
Wij =
1−
dM (Zi , Zj )
dM
max (Zi , Zj )
3 !3
,
(8)
where dM
max (Zi , Zj ) is the largest Mahalanobis distance from the considered observation j
to any observation within the window.
Window. If the choice of the kernel function has a limited impact on the estimation
results (McMillen, 2010),14 the choice of the window is crucial. The common window for
all regressions can be chosen as a minimizer of the mean difference between the dependent variable and the estimated value over all regressions when excluding the considered
observation from the sample.15 The size of the window estimated by cross-validation corre12
Estimating numerous separate regressions for each observation may lead to think that a high degree
of freedom is here used. This is not the case because of the smoothness implied by overlapping samples
(Redfearn (2009), see McMillen and Redfearn (2010) for more details).
13
Compared to the Euclidean distance, the Mahalanobis distance takes into account the correlations of
the data set and is scale-invariant.
14
Results are robust when using a gaussian kernel function instead of a tri-cubic one.
15
The LWR model can be written under a linear form (McMillen and Redfearn, 2010). By denoting
Yi = ln(Pi ), we get
Yi = LYi + ui .
Thus, a cross-validation estimate of h is a minimizer of the cross-validation measure CV (h):
!2
n
1 X Yi − Yˆh,i
,
CV (h) =
n i=1
1 − lii
9
(9)
(10)
sponds to 36% of the initial sample near Bordeaux, 40% near Dunkirk, 76% near Rouen.16
However, larger windows are needed when the objective is to measure marginal effects and
not to predict the dependent variable (McMillen, 2010). This is the reason why we use
windows that correspond to 80% of the initial sample in the three studied cases.17
Thus, locally weighted regression allows coefficients to vary with space, time of sale and
buyer’s characteristics, while keeping some smoothness in their distribution (implied by
overlapping samples). This is why this model enables to relax and to test the assumption of
fixed parameters, that is of constant implicit prices across space, time and among buyers.
Locally weighted regressions have already been used in several works by INSEE. Floch
(2012) provides examples of these works and a methodological presentation of this model
in French.
3
Industrial areas, study areas and data
We have realized an important work of data collection. We detail here the selection of
industrial areas, the delimitation of study areas and the collected variables.
3.1
Industrial areas
Many French hazardous industrial plants are surrounded by a high population density and
could so be considered for this analysis. However, the important work of data collection
limits the number of sites that could be studied. The three industrial areas are chosen
for the different socioeconomic characteristics of their neighborhood and for their different
perceptions of industrial risks.
Socioeconomic characteristics. The three industrial areas studied are located near important French cities: Bordeaux, Dunkirk and Rouen. The three industrial areas present
different socioeconomic characteristics (Table 1). Neighborhoods of the gunpowder near
Bordeaux are the richest ones; many inhabitants are executives in the aerospace induswhere Yˆh,i is the estimate based on the sample from which observation i is removed and using a window
of size h around i.
16
Initial samples include 1,423 observations near Bordeaux, 1,016 near Dunkirk and 571 near Rouen.
Thus, the size of the window estimated by cross-validation corresponds to 512 observations near Bordeaux,
406 near Dunkirk and 434 near Rouen.
17
Here larger windows are all the more needed, as the interest variable (the distance to highly hazardous
plants) varies with some kernel variables (the geographic coordinates): small windows may lead to imprecise
estimates of its marginal effect. However, when considering smaller windows (60%, 40%) with either tricubic or gaussian kernels, the significance and the mean value of the impact of our interest variable on
dwelling value appear as robust.
10
try. Dunkirk and its neighborhoods are the densest area; many households are workers.
Neighborhoods near Rouen are in an intermediate position between the first two areas:
income is higher than near Dunkirk and lower than near Bordeaux; many inhabitants are
executives and work far from their home. In the three areas, according to the 2008 French
annual declaration of social data, at the most 2% of the population in each municipality
is employed by the highly hazardous plants that are described below.
Table 1: Population’s socioeconomic characteristics in the three studied cases
Sports and
sociocultural
facilities (c)
Distance to
labor pool
(d) (km)
26,817
29,005
23,181
41,577
27,599
30,412
14
27
72
12
48
14
5.9
9.2
4.4
11.1
4.8
4.9
11
12
13
16
18,641
20,218
18,610
14,711
23
18
7
14
4.3
4.1
9
6.2
5
1
2
5
1
1
20,083
26,423
21,287
19,638
27,017
23,673
18
2
10
28
5
5
20.5
19.3
20.6
21.1
26.8
20.5
Percentage Average tax
Population
revenue (a)
of built
density (a)
(euro)
area (b)
(per sq km)
Near Bordeaux
Municipality
INSEE
code
Population
(a)
Le Haillan
Martignas-sur-Jalles
Mérignac
Saint-Aubin-de-Médoc
Saint-Médard-en-Jalles
Le Taillan-Médoc
33200
33273
33281
33376
33449
33519
8,378
6,633
65,469
5,550
26,984
8,668
904
251
1,359
159
315
571
Near Dunkirk
8
2
9
1
3
5
Coudekerque-Branche
Dunkirk
Fort-Mardyck
Saint-Pol-sur-Mer
59155
59183
59248
59540
22,994
69,274
3,586
22,100
2,515
1,855
2,543
4,299
Near Rouen
Grand-Couronne
Hautot-sur-Seine
Moulineaux
Petit-Couronne
Sahurs
Val-de-la-Haye
76319
76350
76457
76497
76550
76717
9,346
9,346
881
8,690
1,310
751
552
160
238
679
117
74
Note: Fort-Mardyck and Saint-Pol-sur-Mer municipalities have been associated to Dunkirk
municipality in 2010.
Sources: (a) INSEE, (b) building database of the Geographical National Institute, (c)
topology database of the Geographical National Institute, (d) Sitranet.
Risks and perceptions of industrial activities. Official classification of hazardous
plants is defined by regulation. The Seveso II Directive (Council directive 96/82/EC on the
control of major-accident hazards) defines hazardous industries according to the presence
of hazardous substances or preparations. The degree of hazard is defined with respect to
nature and quantities of substances. Two categories of hazardous facilities are so defined:
“upper tier” sites and “lower tier” sites. The French legislation is actually harsher than
the European one: the French classification includes many other far less hazardous sites,
called “authorized” plants.
The three industrial areas here studied all include hazardous chemical and petrochemical
11
industries. However they present different industrial activities and very different perceptions of associated risks by local populations. The gunpowder factory near Bordeaux was
settled in 1660.18 Today, it comprises two “upper tier” sites of the Seveso II Directive. By
manufacturing gunpowder and explosives, the factory mainly exposes local population to
a risk of explosion. However, it is not necessarily perceived as hazardous by neighboring
population. Indeed, it is a former military plant; only barbed wire can be seen from some
places of the neighborhoods. Furthermore, as the plant is wide (650 buildings over 350
hectares), risk is relatively “contained” within the industrial site. The only nuisance associated to the plant is the transportation of hazardous materials on a precise and limited route.
On the contrary, chemical activities near Dunkirk or Rouen are clearly identified as hazardous by local populations. Near Dunkirk, these activities appeared in the 1970s.19 There
are now sixteen hazardous plants: fourteen “upper tier” Seveso sites, two authorized. Their
activities consist in storage and refining of oil products, metallurgy, manufacture of industrial gases, of chemical and pharmaceutical products and waste treatment. They expose
local population to risks of explosion, fire and toxic impacts. The plants (either buildings,
chimneys or at least plumes of smoke) can be seen from every neighboring dwelling. However, the presence of a nuclear plant in Gravelines (18 km from Dunkirk) may overshadow
the exposure to these industrial risks.
Near Rouen, chemical activities appeared between the 1960s and the 1990s.20 There are
now thirteen hazardous plants: two “upper tier” Seveso sites , one “lower tier” Seveso site,
ten authorized. Their activities are quite diverse: storage and refining of liquid petroleum
gas, production of Diester oil (biodiesel) and of liquid carbon dioxide, perfumes storage,
paper manufacture, goods transportation and warehouses. They expose local population
to the same risks than near Dunkirk: explosion, fire, toxic impacts. However, because of
landscape chimneys or silos cannot be seen from some neighboring dwellings.
Although all these highly hazardous plants were established well before the study period,
several other events, either local or national, may have modified risk perception during this
18
Municipalities in the neighborhoods were initially developed thanks to the activity of the gunpowder.
Urban development was then explained by attraction to Bordeaux center.
19
Near Dunkirk, after World War II, urbanization was realized around industrial activities which were not
perceived as hazardous (shipyards, steel industry); hazardous plants (chemistry, petrochemistry) appeared
in Dunkirk harbor only in the 1970s.
20
Near Rouen, few plants established in the 1920s and 1930s; the majority of plants progressively appeared between the 1960s and the 1990s and cities get closer and closer to these different plants.
12
period, that is between 2000 and 2008. We detail here all these local and national events
(Table 2). Local events are accidents and local policies. Only one accident happened on
January 12, 2007 at Rubis Terminal (storage of liquid flammable and agrochemical products) near Dunkirk.21 Local policies include the distribution of information leaflets, the
update of the emergency plan for households and the implementation of local committees for information and consultation.22 National events are the AZF accident in 2001
in Toulouse and implementation of laws during the study period. The raise of public
awareness of industrial dangers following the AZF accident could have been all the more
important near Rouen that one of the studied plants, Grande Paroisse Normandy, has a
very similar activity to the AZF plant and belongs to the same company, Grande Paroisse
(a subsidiary of Total group).23 Besides, the 2003 law created the technological disasters
insurance system,24 which improves the coverage of households.25
3.2
Delimitation of study areas
In order to measure the impact of distance to plants on housing values, the study area
has to be broad enough to include dwellings close and far from the plants. However, as
Redfearn (2009) notes, “expanding the geographic scope of the sample is likely to draw additional omitted amenities and/or more submarkets”. Indeed, in order to limit the number
of collected characteristics of the dwellings and to avoid districts in which neighborhood
quality is difficult to capture, the central business district and few atypical districts, cor21
Neighboring populations saw flames and plumes of smoke. The accident triggered an emergency plan
inside the plant and required the intervention of civil fire brigades. After the study period, other accidents
occurred. In the gunpowder near Bordeaux, a fire caused one death and two seriously injured persons in
December 2013; near Rouen, mercaptan gas escape from one of the factories, the Lubrizol company, caused
foul odor in France and in England in January 2013.
22
Some future buyers not coming from the study area may not be informed by information releases, in
particular by information leaflets about the emergency plan for households. Indeed, information leaflets
were distributed in 2006 near Dunkirk and in 2007 near Rouen to current owners (Table 2).
23
Indeed, the AZF accident appears to have left a particularly deep and long-lasting impression near
Rouen (“"Grande-Paroisse" à Rouen, un site Seveso hanté par le spectre d’AZF” by Audrey Garric, February 5, 2013. Le Monde.fr ).
24
The 2003 law also created technological risk prevention plans, but they were started after the study
period in each study area. In addition, the very same law created mandatory information for buyers and
tenants: the seller or the landlord has to precise in writing whether his dwelling is located in an area
covered by a technological (or natural) risk prevention plan. This information policy was implemented
on June 1st, 2006. As technological risk prevention plans were started after the study period, no such
information was released in the three studied areas.
25
Technological disasters insurance aims to manage the basic coverage for victims by avoiding long litigation and by covering the residual risk of no responsible identification. Indeed, the guarantee against
technological disasters is mandatorily included in home insurance, which is widely purchased in metropolitan France (Calvet and Grislain-Letrémy, 2011), and the coverage corresponds to the real estate of main
home. Thus, technological disasters premium is included in the home insurance premium; it amounts to
a few euros per year. In case of disaster, households are compensated by their home insurer, which then
turns against the industrialist liable for the damages (or the industrialist’s insurer).
13
Table 2: Local and national events that may have modified industrial risk perception
during the study period (2000-2008)
Year
2001
2002
2003
Bordeaux
-
2004
-
2005
-
2006
-
2007
-
Dunkirk
Information leaflets
Update of the emergency
plan for households
Rouen
-
National level
AZF accident
Law
-
-
-
Creation of local
committee for information
and consultation
-
Creation of local
committee for information
and consultation
+ information leaflets
about the emergency plan
for households
-
Information leaflets
about the emergency plan
for households (∗)
Accident
-
Note: (∗) in all municipalities except Moulineaux.
Source: reports by Technical Studies Center of Public Works of Normandy and Centre, Nord and Picardy,
and South West France.
responding to less than 5% of the samples, are excluded.26
The chosen study areas include dwellings in the very close vicinity of the plants and also
dwellings far from 10 km near Bordeaux, 4 km near Dunkirk and 5 km near Rouen (Table
10 in Appendix A.1 and Figures 1, 2 and 3). The literature studying the impact of similar
industrial risks on housing prices uses similar study areas (Table 13 in Appendix A.2).27
3.3
Data collection
Our analysis requires data on price, characteristics of dwellings, buyers and sellers (Table 3). The price of the dwelling and its intrinsic characteristics (state, living space, number
of rooms, etc.) come from PERVAL notarial data in the years 2000, 2002, 2004, 2006 and
2008. Thus, only dwelling purchases (as opposed to rentings) can be studied.28 Besides,
26
The selection of districts is hereafter detailed for Dunkirk, where this selection is the most important.
Highly attractive districts that are Téteghem municipality (the select residential suburb of Dunkirk), the
historic center and the sandbar in the East of the urban area (Dunkirk-Darses, Dunkirk-Malo-les-Bains)
are excluded. Very unattractive districts such as Mardyck village and Grande Synthe municipalities are
also excluded. Grande Synthe municipality is an urban renewal zone where housing prices are relatively low
because of civil insecurity (petty crime, acts of violence), but this low quality of neighborhood is difficult
to capture. Finally, Leffrinckoucke municipality is excluded, since the presence of a little hazardous plant
but emitting a black plume of smoke can be a source of confusion for inhabitants with the other hazardous
plants here studied.
27
The two exceptions are Carroll et al. (1996), whose study area includes almost 8,000 dwellings, among
which some are 35 km far from the plant, and Boxall et al. (2005), who analyze a rural area and are so led
to study a wide study area to get a sufficient number of observations.
28
Data do not specify whether buyers intend to live in their new dwelling or to rent it.
14
15
Source: base map from the topology database of National Geographic Institute, production by authors.
Figure 1: Study area near Bordeaux
16
Source: base map from the topology database of National Geographic Institute, production by authors.
Figure 2: Study area near Dunkirk
17
Source: base map from the topology database of National Geographic Institute, production by authors.
Figure 3: Study area near Rouen
we consider transactions only between households. Furthermore, we restrict our analysis
to houses, which represent a homogeneous market.29 Each dwelling’s address is precisely
geocoded; detailed data relative to their extrinsic characteristics are collected: commuting
time (by car) to central business district, distance to shops and public utilities, exposure
to industrial risk, to other risks or pollutions.30 Finally, the database from the French
solicitors also provides information about buyers and sellers: their gender, their age, their
social and occupational group, their marital status and their municipality of origin.31 Annual income is allocated using statements of income from the General Directorate of Public
Finances from 2004 to 2008.32
Thus, our database is unique and much richer than the ones used for studies relative
to the impact of similar industrial risks on housing prices, which use data with few (or
without) extrinsic characteristics of the dwellings and without information on buyers and
sellers (Table 13 in Appendix A.2). Our samples finally include 1,423 observations near
Bordeaux, 1,016 near Dunkirk and 571 near Rouen.33 Detailed descriptive statistics are
provided in Appendix A.1 (Tables 9, 10 and 11). The literature studying the impact of
similar industrial risks on housing prices uses a similar number of observations (Table 13
in Appendix A.2).
Our key variable of interest is the distance from the dwelling to highly hazardous plant
(mainly Seveso sites, “upper tier” and “lower tier” ones). We also consider distance to “authorized plants” (subject to the regime of classified plants for the environment protection)
and location in risk management areas. These areas are the emergency planning zone for
households and two areas of control for future land use: the exclusion area, called “Z1
area”, corresponds to the area with lethal damages in case of accident; the area where new
buildings are allowed but limited, called “Z2 area”, corresponds to irreversible damages.
As these zones are very limited, they include a small number of transactions during the
29
Houses are majority in our samples: transactions include 83% of houses near Bordeaux, 87% near
Dunkirk and 72% near Rouen.
30
Geographic distance to central business district was also used and gave similar estimation results.
31
This information is relative to the buyer (or seller) himself, or to the household reference person. Near
Bordeaux, 3% of observations do not include any information on buyers (or only their municipality of
origin) and are excluded from the sample. Near Dunkirk, 6% of observations do not provide the main
information on buyers (gender, age, marital status) and are excluded; the social and occupational group
is missing for 30% of observations. Near Rouen, 9% of observations do not include any information on
buyers (or only their municipality of origin) and are excluded from the sample.
32
Income is allocated using gender, age, marital status and municipality of origin. This leads us to
exclude few foreign buyers (less than 1%) for whom we could not allocate income.
33
In the chosen study areas, transactions within each jurisdiction are relatively uniformly distributed
over the study period (Table 12 in Appendix A.1).
18
study period (Table 10 in Appendix A.1). Finally, as the hazardous plants near Rouen
present the specificity not to be seen from some dwellings, we consider the dummy for view
of plants from the dwelling near Rouen.34
Table 3: Data
Intrinsic characteristics (a)
Price (including tax)
House or apartment
Less than 5 years old
Condition
Living space (?)
Number of rooms
Number of bathrooms
Number of parking lots
Presence of a terrace
Presence of a balcony
Presence of an elevator
Presence of a swimming
pool
Presence of a basement
Presence of a cellar
Presence of annexes
Presence of outbuildings
Presence of an attic
Area of land
Dwelling characteristics
Extrinsic characteristics
Distance and commuting time (by car) to central business
district (∗)
Distance to market square
Distance to drugstore
Distance to food shop
Distance to bus stop
Distance to park
Distance to nursery or primary school
Distance to high school
Distance to the nearest highly hazardous plants (†)
Distance to the nearest authorized plants (‡)
(c),
(c),
(c),
(f)
(f)
(e),
(e),
(h)
(h)
View of industrial plants (near Rouen) ()
(g)
Location in a land use control area (Z1, Z2)
Location in an area of emergency plan for households
Location in an area exposed to natural hazards (.)
Location in an area exposed to other hazards (.)
Location in a residual pollution area
Location in environmental protection area
Location in conservation easement area
Sound exposure to a land transport facility
Sound exposure to an air transport facility
Buyers’ and sellers’ characteristics (/)
(i)
(j)
(i)
(i)
(k)
(i)
(i)
(l), (m)
(l), (m)
(b) ,(f)
(d)
(d)
(d), (e)
(g)
(g)
(n)
Income
(a)
Gender
(a)
Age
(a)
Social and occupational group
(a)
Marital status
(a)
Municipality of origin
Sources: (a) PERVAL, (b) geographic directory of municipalities, (c) Chambers of Commerce and Industry database, (d) municipal database, (e) phone book, (f) topology database of National Geographical
Institute, (g) building database of National Geographical Institute, (h) database for classified plants
per municipality, (i) land use plan, (j) prefecture, (k) Regional Office for Environment, Planning and
Housing, (l) sound map of Departmental Office for Territories and Sea, (m) sound map of Technical
Studies Center of Public Works, (n) statements of income from the General Directorate of Public Finances.
Notes: each distant to a facility is built as the distance to the closest facility.
(∗) Distance or commuting time to central business district is computed as the distance or commuting
time to the town hall of Bordeaux, Dunkirk or Rouen.
(?) Living space is filled in for 81% of observations near Bordeaux, 80% near Dunkirk and 62% near Rouen.
The imputed value for missing values is randomly chosen among the observed distribution of living spaces.
(†) Most hazardous plants among the classified plants for the environment protection (mainly Seveso sites).
(‡) Plants subject to the regime of classified plants for the environment protection.
() View from the dwelling of red and white Petroplus chimney or of Senalia silo.
(.) Area of servitude or notification.
(/) This information is relative to the buyer (or seller) himself, or to the household reference person.
34
The effective distance between the dwelling and the plant is used by all studies relative to the impact
of similar industrial risks on housing prices (Table 13 in Appendix A.2). Other variables are added in some
studies, such as dummies of location in administrative areas for risk management (emergency plans such
as in Boxall et al. (2005) and Travers et al. (2009)) or variables traducing perception of pollutions created
by the plant (Boxall et al. (2005), Sauvage (1997)).
19
4
Results
We estimate the hedonic price function with three different models: the baseline parametric
OLS log-linear model and two extensions of this model, a parametric OLS log-polynomial
model and a semiparametric LWR log-linear model (Section 2). Results of these hedonic
regressions are presented near Bordeaux in Table 4, near Dunkirk in Table 5 and near
Rouen in Table 6.35 As expected, recent building, good condition, living space, number
of rooms, of bathrooms and of parking lots, area of land, all increase the dwelling value.
Presence of outbuildings, of a basement and of a swimming pool can also increase the
dwelling price. As variations of commuting time are limited,36 once including municipality
dummies, the commuting time to central business district does not significantly modify
the dwelling value.37 In the three models, signs are the same and orders of magnitude are
similar.
4.1
Different impacts of distance to highly hazardous plants between
the studied cases
Even among a same category of industries (chemical and petrochemical industries), the impact of distance to highly hazardous plants on the dwelling value varies between the three
cases here studied. Near Bordeaux, in all models, proximity to the gunpowder factory is
valued in the very close vicinity of the gunpowder (Table 4), probably because the neighborhoods of the plant are green with many trees and very quiet places.38 Thus, we capture
here unobserved amenities which are spatially correlated with the distance to the plant.
Near Dunkirk, the semiparametric model reveals that the distance to highly hazardous
plants does not significantly impact housing prices (Table 5).39 Indeed, the presence of
35
Parametric regressions include municipality dummies and year and month dummies. For the sake of
readability, these estimated coefficients are not reported. Some of them are significant at the 1% or the
5% level.
36
All houses are located between a quarter and half an hour from the center, and even between 6 and
14 min near Dunkirk (Table 10 in Appendix A.2).
37
When excluding municipality dummies, commuting time is significant in all parametric models. In
the semiparametric ones, this effect is captured by the distribution of intercepts, as intercept is allowed
to vary with respect to kernel variables, in particular geographic coordinates. All these results are similar
when using the geographic distance to central business district instead of the commuting time.
38
These green areas do not belong to parks. This is why this effect is not captured by the observed
distance to parks.
39
On the contrary, the parametric model indicates that the distance to highly hazardous plants increases
housing prices near Dunkirk. This difference is due, not to the nature of the model used, but to the size of
the sample used for regressions. Near Dunkirk, the main part of the sample is almost uniformly concentrated between 700 m and 3 km from the highly hazardous plants. In the semiparametric regression, the
extreme observations are rarely used and weigh less than the others, whereas in the parametric regression,
the whole sample is used and each observation has an equal weight. When restricting the sample to the
observations between 700 m and 3 km from the highly hazardous plants, the parametric regression shows
that the distance to highly hazardous plants does not significantly impact housing prices.
20
a nuclear plant in Gravelines (located 18 km from Dunkirk) is likely to overshadow the
exposure to industrial risks here studied. Near Rouen, in all models, the distance to highly
hazardous plants increases dwelling values (Table 6): highly hazardous plants are perceived
as disamenities. The magnitude of this impact is important: on average, households are
willing to pay around 1.2% of their dwelling price to go 100 more meters away from highly
hazardous plants near Rouen (Table 7). This order of magnitude is consistent with the
other studies relative to the impact of similar industrial risks on housing prices (Table 13
in Appendix A.2).
Reviewing other studies relative to the impact of similar industrial risks on housing prices
shows that distance to hazardous plants can increase, decrease or have no significant impact on housing prices, depending on the studied case and in particular on the type of
industries (Table 13 in Appendix A.2). Clark and Nieves (1994) also show that the proximity to a petrochemical refinery weighs more on the housing prices than the proximity
to coal/gas/oil-fires plants, to hazardous waste or to a liquefied natural gas site. Our results confirm this dependency. It seems here to be mainly due to the differences in the
perception of industrial activities and of their neighborhoods.
4.2
Testing the fixed-parameter assumption
When comparing the parametric model and the baseline log-linear model, the square or the
cube of distance to highly hazardous plants are significant in the parametric model near
Bordeaux and Rouen (Tables 4 and 6). This suggests that the fixed-parameter assumption does not hold. The fixed-parameter assumption is properly tested by comparing the
semiparametric model and the baseline log-linear model, which corresponds to a nested
model under the null hypothesis of fixed coefficients.40 This hypothesis is rejected at
the 0.01% level, which confirms the need for flexible forms and pleads for the use of this
semiparametric model.
4.3
Comparing the impacts of distance to plants estimated by the parametric and semiparametric models
We compare the estimated impacts of distance to the plants provided by the parametric
log-polynomial model and by the semiparametric model.
40
See McMillen and Redfearn (2010) for a summary of Cleveland and Devlin’s F -test and its adaptation
to the LWR regression.
21
22
Log-linear, OLS
Dwelling price
Estimate Std. error p-value
11.4
0.080
< 2e-16
0.16
0.028
6.1e-09
-0.015
0.014
0.29
-0.21
0.032
4.3e-11
0.0020
2.3e-04
< 2e-16
0.25
0.027
< 2e-16
0.34
0.027
< 2e-16
0.37
0.030
< 2e-16
0.40
0.036
< 2e-16
-0.14
0.042
0.0013
0.13
0.016
7.6e-15
0.099
0.016
1.1e-09
0.15
0.025
1.0e-09
0.052
0.021
0.013
0.12
0.030
2.7e-05
0.13
0.019
1.4e-11
1.6e-05
6.1e-06
0.0092
1.1e-04
6.6e-05
0.11
-2.7e-05
8.9e-06
0.0025
Log-polynomial, OLS
Dwelling price
Estimate Std. error p-value
11.6
0.12
< 2e-16
0.16
0.028
9.1e-09
-0.015
0.014
0.28
-0.22
0.032
2.2e-11
0.0019
2.3e-04
< 2e-16
0.26
0.027
< 2e-16
0.35
0.027
< 2e-16
0.38
0.030
< 2e-16
0.41
0.036
< 2e-16
-0.14
0.042
8.3e-04
0.13
0.016
1.7e-14
0.097
0.016
1.7e-09
0.15
0.025
2.3e-09
0.051
0.021
0.015
0.12
0.030
2.8e-05
0.13
0.019
2.6e-11
2.0e-05
6.5e-06
0.0016
7.5e-05
6.7e-05
0.27
-1.4e-04
5.4e-05
0.012
2.7e-08
1.2e-08
0.026
-1.8e-12
7.8e-13
0.019
Mean
11.3
0.14
-0.032
-0.20
0.0020
0.22
0.31
0.34
0.38
-0.20
0.14
0.078
0.12
-0.034
0.056
0.22
5.0e-05
1.5e-04
-2.0e-05
Log-linear, LWR
Dwelling price
Std. error p-value (∗)
0.32
<1e-04
0.040
<1e-04
0.031
0.077
0.051
<1e-04
4.4e-04
<1e-04
0.056
<1e-04
0.052
<1e-04
0.063
<1e-04
0.071
<1e-04
0.069
0.018
0.030
<1e-04
0.020
0.0014
0.046
3.0e-04
0.023
1.0
0.048
0.029
0.032
<1e-04
4.0e-05
<1e-04
1.7e-04
0.27
2.0e-05
0.022
Notes: parametric regressions include municipality dummies and year and month dummies. These estimated coefficients are not reported. Some of them are significant at the
1% or 5% level.
(†) CBD = Central Business District; (‡) plant denotes a highly hazardous plant.
(∗) See McMillen and Redfearn (2010) for a summary of Cleveland and Devlin’s F -test and its adaptation to test the significance of explanatory variables in the LWR regression.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies Center of Public Works of South West France, statements of income from the
General Directorate of Public Finances. 1,423 observations.
Model
Explained variable
Explanatory variable
Intercept
Less than 5 years
Average condition
Poor condition
Living space
4 rooms
5 rooms
6 rooms
7 rooms or more
No bathroom
2 bathrooms or more
1 parking lot
2 parking lots or more
Presence of outbuildings
Presence of a basement
Presence of a swimming pool
Area of land
Commuting time to CBD (†)
Distance to plants (‡)
(Distance to plants)2
(Distance to plants)3
Table 4: Near Bordeaux: parametric and semiparametric hedonic regressions
23
Log-linear, OLS
Dwelling price
Estimate Std. error p-value
10.9
0.075
< 2e-16
-0.071
0.018
8.8e-05
-0.25
0.021
< 2e-16
0.0011
3.3e-04
6.1e-4
0.15
0.032
1.5e-06
0.24
0.031
1.1e-14
0.30
0.035
< 2e-16
0.32
0.044
7.6e-13
-0.089
0.029
0.0022
0.14
0.035
8.4e-05
-0.038
0.022
0.080
-0.013
0.017
0.45
2.1e-04
3.1e-05
1.1e-10
4.6e-05
1.3e-04
0.72
9.5e-05
2.2e-05
1.6e-05
Mean
10.9
-0.09
-0.23
0.0016
0.12
0.20
0.24
0.24
-0.094
0.13
-0.032
0.022
5.0e-04
7e-05
8e-05
Log-linear, LWR
Dwelling price
Std. error p-value (∗)
0.11
<1e-04
0.023
0.0098
0.052
<1e-04
0.0012
<1e-04
0.030
0.0041
0.041
<1e-04
0.063
<1e-04
0.099
<1e-04
0.044
0.019
0.047
0.013
0.032
0.12
0.030
0.43
2.1e-04
<1e-04
1.8e-04
1.0
2e-05
1.0
Notes: parametric regressions include municipality dummies and year and month dummies. These estimated coefficients are not reported. Some of them are significant at the
1% or 5% level.
(†) CBD = Central Business District; (‡) plant denotes a highly hazardous plant.
(∗) See McMillen and Redfearn (2010) for a summary of Cleveland and Devlin’s F -test and its adaptation to test the significance of explanatory variables in the LWR regression.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies Center of Public Works of Nord and Picardy, statements of income from the General
Directorate of Public Finances. 1,016 observations.
Model
Explained variable
Explanatory variable
Intercept
Average condition
Poor condition
Living space
4 rooms
5 rooms
6 rooms
7 rooms or more
No bathroom
2 bathrooms or more
Presence of outbuildings
Presence of a basement
Area of land
Commuting time to CBD (†)
Distance to plants (‡)
Table 5: Near Dunkirk: parametric and semiparametric hedonic regressions
24
Log-linear, OLS
Dwelling price
Estimate Std. error p-value
11.1
0.21
< 2e-16
-0.061
0.026
0.020
-0.31
0.042
2.2e-13
0.0017
3.6e-04
3.7e-06
0.22
0.040
2.9e-08
0.34
0.039
< 2e-16
0.40
0.044
< 2e-16
0.47
0.049
< 2e-16
0.14
0.030
6.0e-06
0.077
0.030
0.010
0.12
0.035
5e-04
0.040
0.029
0.17
0.098
0.024
3.5e-05
2.4e-05
9.3e-06
0.010
-1.5e-04
1.3e-04
0.25
1.3e-04
2.1e-05
3.7e-10
Log-polynomial, OLS
Dwelling price
Estimate Std. error p-value
10.8
0.23
< 2e-16
-0.059
0.026
0.025
-0.31
0.041
2.7e-13
0.0018
3.6e-04
2.0e-06
0.21
0.039
1.2e-07
0.34
0.039
< 2e-16
0.39
0.043
< 2e-16
0.46
0.048
< 2e-16
0.14
0.030
4.9e-06
0.064
0.030
0.030
0.11
0.035
0.0017
0.042
0.029
0.15
0.092
0.023
1.0e-04
2.5e-05
9.2e-06
0.0074
-6.5e-06
1.4e-04
0.96
2.6e-04
4.7e-05
4.4e-08
-3.3e-08
1.1e-08
0.0021
Mean
11.0
0.028
-0.29
0.0019
0.22
0.36
0.34
0.43
0.087
0.10
0.11
-0.014
0.078
8.0e-05
-1.1e-04
1.2e-04
Log-linear, LWR
Dwelling price
Std. error p-value (∗)
0.27
<1e-04
0.040
0.93
0.068
<1e-04
5.8e-04
0.0011
0.055
<1e-04
0.071
<1e-04
0.052
<1e-04
0.065
<1e-04
0.059
0.0028
0.038
0.29
0.063
0.046
0.035
0.59
0.024
0.0077
5.0e-05
<1e-04
1.7e-04
1.0
5.0e-05
<1e-04
Notes: parametric regressions include municipality dummies and year and month dummies. These estimated coefficients are not reported. Some of them are significant at the
1% or 5% level.
(†) CBD = Central Business District; (‡) plant denotes a highly hazardous plant.
(∗) See McMillen and Redfearn (2010) for a summary of Cleveland and Devlin’s F -test and its adaptation to test the significance of explanatory variables in the LWR regression.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies Center of Public Works of Normandy and Centre, statements of income from the
General Directorate of Public Finances. 571 observations.
Model
Explained variable
Explanatory variable
Intercept
Average condition
Poor condition
Living space
4 rooms
5 rooms
6 rooms
7 rooms or more
2 bathrooms or more
1 parking lot
2 parking lots or more
Presence of outbuildings
Presence of a basement
Area of land
Commuting time to CBD (†)
Distance to plants (‡)
(Distance to plants)2
Table 6: Near Rouen: parametric and semiparametric hedonic regressions
Value. Parametric log-polynomial regressions lead to a bias in the estimated value of
the impact of distance to highly hazardous plants on dwelling prices. Near Bordeaux
and Rouen, this implicit price of distance d to these plants equals the absolute marginal
willingness to pay AM W T P to go one more meter away from highly hazardous plants
(Section 2). To make comparisons between dwellings or between the three cases easier, we
also compute the relative willingness to pay RM W T P to go further away from the plants,
which expresses this willingness to pay as a percentage of the price of the considered
dwellings to go 100 more meters away from the plants.
AM W T P =
∂P (X)
∂d ,
RM W T P =
∂P (X)/∂d
.
P
Table 7 properly compares the absolute and relative marginal willingnesses to pay, as estimated by the parametric and semiparametric models.
Table 7: Households’ marginal willingness to pay to go one more meter away from highly
hazardous plants
MWTP estimated by
log-polynomial, OLS
Median Mean Std. Dev.
Near Bordeaux
AMWTP (=
C/m)
RMWTP (%/100m)
Near Dunkirk
AMWTP (=
C/m)
RMWTP (%/100m)
Near Rouen
AMWTP (=
C/m)
RMWTP (%/100m)
MWTP estimated by
log-linear, LWR
Median Mean Std. Dev.
-3.65
-0.19
-5.32
-0.25
5.24
0.21
-3.36
-0.19
-3.17
-0.20
3.89
0.21
9.45
0.95
10.00
0.95
3.91
0
-
-
-
19.33
1.88
20.25
1.67
11.16
0.68
14.39
1.22
15.44
1.20
9.91
0.51
Notes: in the parametric regression, the log of price is explained by the linear expression of dwelling
characteristics and by a polynomial of the distance to the plants (Equation 4). In the semiparametric
regression, the log of price is explained by the linear expression of dwelling characteristics while allowing
the coefficients to vary with respect to Zi , that is with respect to the geographic coordinates, time and the
buyer’s income (Equation 6). Thus, we get
Log-polynomial OLS Bordeaux:
Log-polynomial OLS Dunkirk:
Log-polynomial OLS Rouen:
Log-linear LWR:
AM W T P
AM W T P
AM W T P
AM W T P
= (β1 + 2β2 d + 3β3 d2 )P (X),
= β1 P (X),
= (β1 + 2β2 d)P (X),
= β(Zi )P (X),
RM W T P
RM W T P
RM W T P
RM W T P
= (β1 + 2β2 d + 3β3 d2 ),
= β1 ,
= (β1 + 2β2 d),
= β(Zi ).
Near Dunkirk, willingnesses to pay as estimated by the LWR are not provided, as the distance to the
nearest highly hazardous plant is not significant in this model.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies Center of
Public Works of Normandy and Centre, statements of income from the General Directorate of Public
Finances. 1,423 observations near Bordeaux, 1,016 near Dunkirk and 571 near Rouen.
The parametric model leads to a biased estimation of the average marginal willingness to
pay over the sample, be it the absolute willingness or the relative one, near Bordeaux and
25
Rouen. Near Rouen, it leads to a clear overestimation: the mean of the absolute marginal
willingness to pay and the one of the relative marginal willingness to pay are 31% and 39%
respectively higher as estimated by the parametric model than by the semiparametric one
(Table 7). It is worth underlining that the bias may lead either to an overestimation or to
an underestimation of the average marginal willingness to pay. Indeed, the semiparametric
model gives a higher weight to sets of “close” observations (proximity being defined by the
kernel in terms of location, time and income), whereas the parametric model gives an equal
weight to each observation.41 If, as here, these sets of “close” observations correspond to
a lower marginal willingness to pay than the average, the parametric model leads to an
overestimation of the average marginal willingness to pay.
Variations. Near Rouen, the parametric model also leads to a biased estimation (here
again a clear overestimation) of the variations of the marginal willingness to pay with respect to the distance to plants.42
We perform here regressions that correspond to the second step of Rosen (1974)’s method.
The first step is the main hedonic price regression: it estimates the dwelling price P (X)
depending of the dwelling attributes X (Tables 4, 5 and 6 in Section 4). The second step
here conducted consists in using the implicit price (∂P (X)/∂xk )k estimated in the first
step, by either the parametric or the semiparametric model, to recover information relative to demand and supply. More precisely, we explain the marginal willingness to pay
as estimated by the semiparametric regression with the distance to plants, year dummies,
buyer’s income and also his age, gender, social and occupational group, marital status and
location of origin (Table 8). We explain the marginal willingness to pay as estimated by
the log-polynomial parametric regression with the distance to plants (Table 8).43
These regressions must be considered as essentially descriptive. Indeed, the marginal price
of each dwelling’s characteristic may vary with the quantity of this characteristic. Thus,
households choose simultaneously the quantity and the marginal price for each characteristic. This creates an identification problem in this regression as we explain the willingness
41
Indeed, OLS can be considered as a special case of LWR: it is a LWR with a windowsize of 100% (the
whole sample) and a weight equal to 1/n everywhere, n denoting the sample size.
42
We describe here the relative willingness to pay to go further away from highly hazardous plants.
Results are similar when using the absolute willingness to pay.
43
Indeed, the log-polynomial parametric model does not allow the marginal willingness to pay to vary
with respect to time or buyer’s income. When explaining the marginal willingness to pay as estimated by
the parametric regression also with respect to time of sale and all buyer’s characteristics, we find only a
significant impact of time (a significant increase in 2002, 2004 and 2006).
26
to pay to go away from the plants (the marginal price of distance) with the distance to
these plants.44
Table 8: Explaining households’ relative marginal willingness to pay to go one more meter
away from highly hazardous plants near Rouen
RMWTP estimated
by log-linear, LWR
Explanatory variable
Estimate Std. error p-value
Intercept
1.46
0.16
< 2e-16
Distance to plants
-2.4e-04
1.8e-05
< 2e-16
2002
0.089
0.054
0.10
2004
0.11
0.057
0.046
2006
0.26
0.060
3.1e-05
2008
0.38
0.058
1.7e-10
Income
-1.3e-07
1.2e-06
0.91
Age
1.0e-04
2.2e-03
0.96
Gender (men)
0.064
0.052
0.22
Farmer
-0.043
0.31
0.89
Self-employed
-0.092
0.11
0.39
Laborer
-0.045
0.091
0.62
Manager or professional
-0.060
0.094
0.52
Intermediate lower occupation
-0.049
0.093
0.60
Intermediate upper occupation
-0.062
0.086
0.47
Single
-0.16
0.048
7.4e-4
From more than 15 km (France)
-0.050
0.068
0.46
From abroad
-0.080
0.43
0.85
Note: buyer’s characteristics are relative to the buyer himself, or to the household reference person.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies Center of
Public Works of Normandy and Centre, statements of income from the General Directorate of Public
Finances. 571 observations.
Explained variable
RMWTP estimated by
log-polynomial, OLS
Estimate Std. error p-value
2.6
5.7e-17
<2e-16
-6.7e-04
3.2e-20
<2e-16
Households’ willingness to pay to go further away from highly hazardous plants decreases
with respect to the distance to plants over the study area near Rouen (Table 8, Figure 4).
This is due to the fact that marginal gains (in terms of exposure reduction) of going further away from the hazard source are likely to decrease. However, two other phenomenons
may explain the variations of willingness to pay with respect to the distance to plants.
First, other spatially-correlated amenities could bias these estimated variations. Second,
there may be a sorting on housing market, uncaptured by buyer’s observed characteristics:
households with a higher willingness to pay to live further away from the plants may precisely choose a more distant location from these plants. We observe the net result of all
three effects; it appears that the first effect dominates the last two ones. The decrease of
households’ willingness to pay to go further away with respect to the distance to plants
is overestimated by the parametric model: corresponding coefficient is threefold higher
44
Ekeland et al. (2004) offer two methods to implement this second step regression using data from a
single market. However, our data do not enable us to apply them, as variability of dwelling attributes with
respect to buyers’ observed characteristics is required. This simultaneous choice also creates an endogeneity
bias in the first step.
27
(Table 8, Figure 4).
2.0
1.5
1.0
0.5
0.0
●
●● ●
●
●
●
●
● ● ● ●●
●
●
●
●
●
●
●
●● ●●
●●
●●
●●
●
● ● ●
●●
●●●
●
●
●
●● ●●●
●
●
●●
●
●● ● ●●●
●
●
●● ● ●
●●●● ●
●
●
●● ●
●
●
●
●
●
●
●
●
●
●
●● ●
●●● ●
● ●
●●● ●●
●
●
●
●
●
●●● ●
● ●
●
●● ●
●●● ● ● ●●●
●
●
● ●●
● ●
●● ●● ●
●
● ● ●
●●●●●
●
●●
●
●
●
●
●●
●●
●
● ●
●●
●
● ● ●●
●
●
●
●
●
●
● ●●
● ●●●
●
● ●●
●●
●●
●● ●
●
●
●
● ●
●
●
●
●
●●●
●
●●
●
● ●●
●
●●
●
●●
●
●
●●
●
●●
●
●
●
●●●
●
●
●
●
●●
●
●
●
●● ● ● ●● ●
●● ●
● ●●●
●●●
● ●
●
●
● ● ●● ●
●●
● ●
●
●
●
●
● ●
●
●
●
● ●
● ● ●
●●
●
●●
●
●
●●
●
● ● ●●●●●
●●
● ●
●
●
●●●●● ● ●
●●●
●
●●●
●●●
●●
●
● ● ●
● ●
●●●
●
● ●● ●●
●
●●●●
●
●
●●●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●●
●●
● ● ●●
●
●●
●
●●
●
●●
●●
●
●
●
●●
●
●
●
● ●
●
●
●●
●
●● ●●
●
●
●●●
●
●
●
●
●●●
● ● ●●
●
●●●
●
●
●
●
●●●●●
●
●
●●
● ●
●
●●
●
●
●
●
●
●
●
●
●
●
● ●
●
●●
●
●
●
●
●
●
●
●
●
●●●
● ● ●
●
●
●●
●●
●● ●
●
●
●
●
●
●
●●● ●●
●
● ● ●
●●● ●
●
●
●
●
●
●●
● ●●
● ●
● ●
● ●
●
●
●●
●
●
●
● ●
● ●
●
●
●
●●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
● ●
●
−0.5
Relative willingness to pay (%/100m)
2.5
Figure 4: Households’ marginal willingness to pay to go one more meter away from highly
hazardous plants near Rouen
0
1000
2000
3000
4000
5000
Distance to highly hazardous plants (m)
Caption: The relative marginal willingness to pay to go 100 more meters away from highly
hazardous plants, as estimated from the semiparametric model, is a scatter plot. The slope
as estimated by the semiparametric model is represented by a dark straight line. The one
estimated from the log-polynomial parametric model is represented by a grey straight line.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies
Center of Public Works of Normandy and Centre, statements of income from the General
Directorate of Public Finances. 571 observations.
In the semiparametric model, households’ willingness to pay to go further away from highly
hazardous plants increases over time following the 2001 AZF accident,45 the 2003 law,46 or
2005 and 2007 information policies (Table 8). Buyer’s characteristics that are income, age,
gender, social and occupational group do not significantly explain the willingness to pay
to go further away from the plants (Table 8).47 The single have a lower relative willingness
45
Recall that Grande Paroisse Normandy, which is settled in near Rouen, has a very similar activity to
the AZF plant and belongs to the same company, Grande Paroisse (a subsidiary of Total group). Another
analysis of the impact of the AZF accident on housing prices near a French facility similar to the AZF
plant (Travers et al., 2009) (Port-Jérôme harbor, Seine-Maritime, France) shows the absence of impact of
the AZF accident, while using a parametric model.
46
In the short term, this law implemented the technological disasters insurance system, which improves
the coverage of households against technological disasters and should have decreased the marginal willingness to pay. However, even if they were not straightaway effective, other measures of this law with
a negative impact on homeowners (such as the implementation of technological risk prevention plans or
mandatory information about risks) have been probably more mediatized.
47
This absence of significant impact holds when considering other specifications (with polynomials or
log).
28
to pay to further away from highly hazardous plants, probably as they have no children.
Buyers coming from less than 15 km of the study area have a similar willingness to pay to
go further away from highly hazardous plants.
5
Conclusion
Real estate markets can reveal households’ willingness to pay to reduce their exposure to
hazardous industrial facilities. With highly detailed microdata, we study housing prices in
the vicinity of hazardous industries located near three important French cities, Bordeaux,
Dunkirk and Rouen. We estimate the impact of distance to hazardous plants on dwelling
values by using hedonic price models.
We compare the results from standard parametric hedonic property models and an alternative, more flexible, semiparametric hedonic property model. This semiparametric model
is a locally weighted regression that allows this willingness to pay to vary with respect to
space, buyers’ characteristics and time, while keeping some smoothness in their distribution. If signs and orders of magnitude of effects are similar in the two models for the very
wide majority of coefficients, the estimated impacts of the distance to highly hazardous
plants on housing prices significantly differ between the two models. The parametric model
leads to an important bias in the estimated value of this impact near Bordeaux and Rouen
and in its variations with respect to distance to the plants near Rouen.
Using the semiparametric model, we show that this impact strongly differs among industrial areas, even among chemical and petrochemical industries, likely because of different
perceptions of industrial risks and dissimilar neighborhoods of these hazardous facilities.
We also show that this impact may vary within a study area. Indeed, the impact of distance to hazardous plants on housing prices may significantly decrease with respect to the
distance to the plants, as marginal gains in terms of exposure reduction of going further
away from the hazard source are likely to decrease; it may also vary over time following
accidents or information policies.
There are two substantive lessons that one learns from our analysis. First, our results
show that parametric models can lead to an important bias in the estimated value of the
marginal willingness to pay. Second, our results show that estimated willingness to pay for
prevention strongly differs among industrial areas, even among one category of industries
29
(here chemical and petrochemical industries), and depends on the distance to these facilities
and on time. This inadequate estimation method and this limited external validity plead
for a careful use of current estimations of population’s willingness to pay for industrial risk
reduction in the cost-benefit analyses of prevention measures, whereas most cost-benefit
analyses have until now used the parameters taken from the estimation of parametric
hedonic models on other study area and period than the ones under consideration.
References
Anglin, P.M., Gencay, R., 1996. Semiparametric Estimation of a Hedonic Price Function.
Journal of Applied Econometrics 11, 633–648.
Bin, O., Kruse, J.B., Landry, C.E., 2008. Flood Hazards, Insurance Rates, and Amenities:
Evidence From the Coastal Housing Market. Journal of Risk and Insurance 75, 63–82.
Bontemps, C., Simioni, M., Surry, Y., 2008. Semiparametric Hedonic Price Models: Assessing the Effects of Agricultural Nonpoint Source Pollution. Journal of Applied Econometrics 23, 825–842.
Boxall, P.C., Chan, W.H., McMillan, M.L., 2005. The Impact of Oil and Natural Gas
Facilities on Rural Residential Property Values: A Spatial Hedonic Analysis. Resource
and Energy Economics 27, 248–269.
Calvet, L., Grislain-Letrémy, C., 2011. L’Assurance Habitation dans les Départements
d’Outre-mer : Une Faible Souscription. Economie et Statistique Numéro 447, 57–70.
Carroll, T.M., Clauretie, T., Jensen, J., Waddoups, M., 1996. The Economic Impact of
a Transient Hazard on Property Values: The 1988 PEPCON Explosion in Henderson,
Nevada. Journal of Real Estate Finance and Economics 13, 143–167.
Chay, K.Y., Greenstone, M., 2005. Does Air Quality Matter? Evidence From the Housing
Market. Journal of Political Economy 113, 376–424.
Clark, D.E., Nieves, L.A., 1994. An Interregional Hedonic Analysis of Noxious Facility
Impacts on Local Wages and Property Values. Journal of Environmental Economics and
Management 27, 235–253.
Currie, J., Davis, L.W., Greenstone, M., Walker, R., 2013. Do Housing Prices Reflect
Environmental Health Risks? Evidence from More than 1600 Toxic Plant Openings and
Closings. National Bureau of Economic Research .
Davis, L.W., 2011. The Effect of Power Plants on Local Housing Values and Rents. Review
of Economics and Statistics 93, 1391–1402.
Ekeland, I., Heckman, J.J., Nesheim, L., 2004. Identification and Estimation of Hedonic
Models. Journal of Political Economy 112, 60–109.
Farber, S., 1998. SURVEY Undesirable Facilities and Property Values: A Summary of
Empirical Studies. Ecological Economics 24, 1–14.
30
Floch, J.M., 2012. Détection des disparités socio-économiques. L’apport de la statistique
spatiale. Technical Report H 2012/04. INSEE. Direction de la Diffusion et de l’Action
régionale.
Flower, P.C., Ragas, W.R., 1994. The Effects of Refineries on Neighborhood Property
Values. Journal of Real Estate Research 9, 319–338.
Freeman, A.M., 2003. The Measurement of Environmental and Resource Values: Theory
and Methods. Resource for the Future. chapter Property Value Models.
Gayer, T., Hamilton, J.T., Viscusi, W.K., 2000. Private Values of Risk Tradeoffs at Superfund Sites : Housing Market Evidence on Learning about Risk. Review of Economics
and Statistics 82, 439–451.
Greenstone, M., Gallagher, J., 2008. Does Hazardous Waste Matter? Evidence From
the Housing Market and the Superfund Program. Quarterly Journal of Economics 123,
951–1003.
Greenstone, M., Hornbeck, R., Moretti, E., 2010. Identifying Agglomeration Spillovers:
Evidence from Winners and Losers of Large Plant Openings. Journal of Political Economy 118, 536–598.
Grislain-Letrémy, C., Katossky, A., 2013. L’Impact de l’Industrie à Hauts Risques sur le
Prix des Logements. Economie et Statistique 460-461, 79–106.
Harrison, D.M., Smersh, G.T., Arthur L. Scwartz, J., 2001. Environmental Determinants
of Housing Prices : The Impact of Flood Zone Status. Journal of Real Estate Research
21.
Kohlhase, J.E., 1991. The Impact of Toxic Waste Sites on Housing Values. Journal of
Urban Economics 30, 1–26.
Maani, S.A., 1991. Risk and Information: A Hedonic Price Study in the New Zealand
Housing Market. Economic Record 67, 227–236.
MacDonald, D.N., White, H.L., Taube, P.M., Huth, W.L., 1990. Flood Hazard Pricing
and Insurance Premium Differentials: Evidence from the Housing Market. Journal of
Risk and Insurance 57, 654–663.
Maslianskaïa-Pautrel, M., 2008. La Valorisation de la Qualité de l’Air Par l’Approche
Hédonique : Une Revue de la Littérature .
McMillen, D.P., 2010. Issues in Spatial Data Analysis. Journal of Regional Science 50,
119–141.
McMillen, D.P., Redfearn, C.L., 2010. Estimation and Hypothesis Testing for Nonparametric Hedonic House Price Functions. Journal of Regional Science 50.
Morgan, A., 2007. The Impact of Hurricane Ivan on Expected Flood Losses, Perceived
Flood Risk, and Property Values. Journal of Housing Research 16, 47–60.
Pope, J., 2008. Buyer Information and the Hedonic: The Impact of a Seller Disclosure on
the Implicit Price for Airport Noise. Journal of Urban Economics, Volume 63, Issue 2,
Pages 498-516 .
Redfearn, C.L., 2009. How Informative are Average Effects? Hedonic Regression and
Amenity Capitalization in Complex Urban Housing Markets. Regional Science and
Urban Economics 39, 297–306.
31
Rosen, S., 1974. Hedonic Prices and Implicit Markets: Product Differentiation in Pure
Competition. Journal of Political Economy 82, 34–55.
Sauvage, L., 1997. L’Impact du Risque Industriel sur l’Immobilier. Association des Etudes
Foncières.
Sunding, D.L., Swoboda, A.M., 2010. Hedonic Analysis with Locally Weighted Regression: An Application to the Shadow Cost of Housing Regulation in Southern California.
Regional Science and Urban Economics 40, 550–573.
Travers, M., Bonnet, E., Chevé, M., Appéré, G., 2009. Risques Industriels et Zone Naturelle
Estuarienne : Une Analyse Hédoniste Spatiale. Economie et Prévision 4-5, 136–158.
Treich, N., 2005. L’Analyse Coût-Bénéfice de la Prévention des Risques. Technical Report.
LERNA-INRA, Université de Toulouse.
A
A.1
Appendices
Descriptive statistics
32
Table 9: Intrinsic characteristics of the dwellings
Bordeaux Dunkirk Rouen
N
1423
1016
571
Price (current euro, tax. inc.) Min
22090
7622
15245
Q1
146428
76222
92994
Median
197000
99092 115850
Q3
256305
128000 154000
Max
800000
285300 430000
Mean
206887
104860 126402
Standard deviation
85661
41021
54707
Less than 5 years old
83
6
15
Condition
Good
497
525
179
Average
857
297
339
Poor
69
194
53
Living space (sq m) (∗)
Min
30
46
30
Q1
95
85
83
Median
110
95
100
Q3
134
110
120
Max
350
300
300
Mean
119
100
104
Standard deviation
36
25
34
Unspecified
260
210
268
Number of rooms
3 or less (†)
109
77
71
4
361
268
131
5
506
441
177
6
288
161
115
7 or more
159
69
77
Number of bathrooms
0
34
80
0
1
917
881
450
2 or more
472
55
121
Number of parking lots
0
304
1016
127
1
971
0
299
2 or more
148
0
145
Holds...
... a terrace
83
9
56
... a swimming pool
207
0
0
... a basement
72
345
266
... a cellar
83
47
255
... outbuildings
161
154
118
... an attic
70
65
202
Area of land (sq m)
Min
30
0
30
Q1
606
155
316
Median
816
207
500
Q3
1052
305
848
Max
29597
6599
18724
Mean
967
266
783
Standard deviation
1115
261
1355
Notes: the unit is the number of dwellings, unless otherwise specified.
(∗) The imputed living space for missing values is randomly chosen among the observed
distribution of living spaces.
(†) Observations with zero room (stand-alone garages, garden sheds or other outbuildings)
are excluded.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies
Center of Public Works of Normandy and 33
Centre, Nord and Picardy, and South West
France.
Table 10: Extrinsic characteristics of the dwellings
Bordeaux Dunkirk Rouen
N
1423
1016
571
Min
16
6
16
Commuting time
Q1
24
8
19
(by car) to city
Median
28
9
22
center (min)
Q3
31
10
25
Max
38
14
31
Mean
27
9
22
Standard deviation
4
2
4
Located close to...
... market square (<500m)
205
295
97
... a drugstore (<250m)
144
460
109
... a food shop (<250m)
198
686
214
... a bus stop (<250m)
676
880
448
... a park (<500m)
425
714
119
... a nursery or a primary school (<500m)
428
924
172
... a high school (<500m)
158
541
161
Min
532
41
64
Distance to the
Q1
3333
837
705
nearest highly
Median
5141
1730
1140
hazardous plant
Q3
6406
2451
2189
(m)
Max
10379
4084
5142
Mean
4935
1711
1452
Standard deviation
1863
949
1027
Min
89
41
64
Distance to the
Q1
1006
596
494
nearest
Median
1657
872
771
authorized plant
Q3
2663
1163
1189
(m)
Max
5279
1943
2884
Mean
1848
877
943
Standard deviation
1061
369
624
View of industrial plants
0
0
495
Located in an
... land use control (Z1)
0
0
34
administrative
... land use control (Z2)
0
0
56
area of...
... emergency plan
0
276
391
... natural hazard
33
0
0
... other hazard
0
0
58
... environmental protection
0
0
61
... conservation easement
0
39
200
Exposed to...
... land transport noise (?)
0
0
15
... air transport noise (?)
79
0
0
Notes: the unit is the number of dwellings, unless otherwise specified.
(?) A dwelling is considered as exposed to a land transport facility if sound is above 60 dB / to an air
transport facility if sound is above 50 dB.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies Center of Public
Works of Normandy and Centre, Nord and Picardy, and South West France.
34
Table 11: Buyers’ characteristics
Bordeaux Dunkirk Rouen
N
1423
1016
571
Income (euro)
Min
5024
3471
7550
Q1
20498
12134
13563
Median
38815
23266
28958
Q3
47942
32544
35408
Max
173016
160996 277797
Mean
35609
23781
27161
Standard deviation
16061
12348
17777
Unspecified
112
64
114
Gender
Female
229
200
99
Male
1194
816
472
Age (year)
Min
20
20
22
Q1
34
29
31
Median
40
34
37
Q3
48
43
46
Max
99
82
85
Mean
42
37
40
Standard deviation
11
11
11
Unspecified
0
0
1
Farmer
3
1
2
Social and
Self-employed
71
31
31
occupational
Managerial or professional occupation
445
47
77
group
Intermediate upper occupation
472
203
188
Employee
193
112
91
Worker
103
180
131
Retired
100
28
41
No occupation
22
13
7
Unspecified
14
401
3
Marital status
Single
301
322
172
In civil union
16
17
7
Married
884
538
309
Divorced
158
101
50
Widowed
34
18
18
Originates from... ... study area
633
758
199
... less than 15 km away
358
155
324
... more than 15 km away (France)
391
95
47
... abroad
11
6
1
Unspecified
30
2
0
Notes: buyer’s characteristics are relative to the buyer himself, or to the household reference
person. The unit is the number of buyers, unless otherwise specified.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies
Center of Public Works of Normandy and Centre, Nord and Picardy, and South West
France, statements of income from the General Directorate of Public Finances.
35
Table 12: Transactions by municipality and by year
Municipality (INSEE code)
2000 2002 2004 2006 2008 All years
Le Haillan (33200)
36
31
34
38
24
163
Martignas-sur-Jalle (33273)
42
40
42
32
30
186
Mérignac (33281)
13
15
22
13
11
74
Saint-Aubin-de-Médoc (33376)
34
38
38
28
21
159
Saint-Médard-en-Jalles (33449)
48
157
155
149
84
593
Le Taillan-Médoc (33519)
52
50
55
49
42
248
All municipalities
225
331
346
309
212
1423
Dunkirk
Coudekerque-Branche (59155)
80
77
87
86
85
415
Dunkirk (59183)
47
35
49
39
41
211
Fort-Mardyck (59248)
9
12
8
17
16
62
Saint-Pol-sur-Mer (59540)
50
67
86
56
69
328
All municipalities
186
191
230
198
211
1016
Rouen
Grand-Couronne (76319)
48
58
57
42
50
255
Moulineaux (76457)
6
10
4
7
5
32
Petit-Couronne (76497)
43
45
29
28
38
183
Sahurs (76550)
10
15
9
9
5
48
Val-de-la-Haye (76717)
7
7
13
8
6
41
All municipalities
118
140
113
94
106
571
All areas
529
662
689
601
529
3010
Note: the unit is the number of transactions.
Sources: French solicitors - PERVAL, data collected and standardized by Technical Studies
Center of Public Works of Normandy and Centre, Nord and Picardy, and South West
France.
Area
Bordeaux
A.2
Other studies on industrial risk and housing prices
36
Chemical and
petrochemical
facilities
Bradford (United
Kingdom)
Henderson,
Nevada (United
States)
1987-1993
1986 - 1990
561
170, 64, 59,
91 and 188
1,999
7,780
532
Sample
size
No
No
No
No
Distance to the
City, view of
mountains
Extrinsic
characteristics (*)
From 0 to 2.5
km
From 0 to 1.8
km/2
km/4 km
From 0 to 3.3
km
From 3 to 35
km
17 km
(mean), 7 km
(std) (rural
area)
Scope of the
study area
(**)
Distance to the plants, view of
plants
Distance to the plants
Distances to the refineries or
dummies for areas based on
these distances
Distance to the plant
Prices decrease for
properties located within 4
km of facilities from
150,000 to 450,000 $CDN
2001, i.e. from -4% to -8%
Distance to the nearest plant,
number of plants within 4 km,
number of emergency planning
zones the property is located
in, sum of H2 S release within 4
km of property
Travers
et al.
(2009)
Chemical
facility
Port-Jérôme, Seine
Maritime (France)
Prices increase by =
C9.2 per m,
i.e. by 1.2% per 100 m
Prices increase by £5.5 per m,
i.e. around 2.6% per 100 meter.
View over facilities decreases
prices by £960, i.e. by 4.6%
In Waziers and in
Puget-sur-Argens, no
significant impact. Elsewhere,
prices increase w.r.t. distance.
In L’Hôpital for example, prices
increase by 110 F per meter,
i.e. around 4% per 100 meter
In one area prices decrease
w.r.t distance. Elsewhere prices
increase from $210 to $620 per
100 m, i.e. from 1% to 3.4% per
100 m, i.e. from $2 to $6 per m
Prior to explosion, prices
increase by 4.17% at 3.2 km
from PEPCON, i.e. increase by
around $11 per meter. Impact
of the explosion on implicit
price
Effect of risk exposure
Variable(s) for risk exposure
Distances to roads,
highway, Seine,
From 0 to 5
Distance to the plants, location
2001-2002
228
play area, town
km
in emergency planning zone
hall
(*) Other than variables for risk exposure and dummies for municipalities. (**) The scope of the study area is measured in terms of distance to the nearest plant.
Chemical
facilities
1988-1992
Waziers,
Puget-sur-Argens,
Carling, L’Hôpital
and St-Gaudens
(France)
Petrochemical
refineries
Flower
and Ragas
(1994)
Sauvage
(1997)
1979 - 1991
St. Bernard
Parish, Louisiana
(United States)
PEPCON
plant
(chemistry)
1994 - 2001
Carroll
et al.
(1996)
Central Alberta
(Canada)
Oil and gas
facilities
Time
Boxall
et al.
(2005)
Region
Hazardous
facilities
Study
Table 13: Other studies on industrial risk and housing prices
Liste des documents de travail de la Direction des Études et Synthèses Économiques
G 9001
J. FAYOLLE et M. FLEURBAEY
Accumulation, profitabilité et endettement des
entreprises
G 9002
H. ROUSSE
Détection et effets de la multicolinéarité dans les
modèles linéaires ordinaires - Un prolongement
de la réflexion de BELSLEY, KUH et WELSCH
G 9003
P. RALLE et J. TOUJAS-BERNATE
Indexation des salaires : la rupture de 1983
G 9004
D. GUELLEC et P. RALLE
Compétitivité, croissance et innovation de produit
G 9005
P. RALLE et J. TOUJAS-BERNATE
Les conséquences de la désindexation. Analyse
dans une maquette prix-salaires
G 9101
ii
Macro-economic import functions with imperfect
competition - An application to the E.C. Trade
G 9203
G 9204
G 9205
I. STAPIC
Les échanges internationaux de services de la
France dans le cadre des négociations multilatérales du GATT
Juin 1992 (1ère version)
Novembre 1992 (version finale)
P. SEVESTRE
L'économétrie sur données individuellestemporelles. Une note introductive
françaises : une évaluation empirique des théories de la structure optimale du capital
Équipes Amadeus (INSEE), Banque de France,
Métric (DP)
Présentation des propriétés des principaux modèles macroéconomiques du Service Public
G 9414
G 9314
B. CREPON - E. DUGUET
Research & Development, competition and
innovation
I. KABLA
Le Choix de breveter une invention
G 9501
B. DORMONT
Quelle est l'influence du coût du travail sur
l'emploi ?
J. BOURDIEU - B. CŒURÉ - B. SEDILLOT
Irreversible Investment and Uncertainty:
When is there a Value of Waiting?
G 9502
L. BLOCH - B. CŒURÉ
Imperfections du marché du crédit, investissement des entreprises et cycle économique
G 9503
D. GOUX - E. MAURIN
Les transformations de la demande de travail par
qualification en France
Une étude sur la période 1970-1993
G 9504
N. GREENAN
Technologie, changement organisationnel, qualifications et emploi : une étude empirique sur
l'industrie manufacturière
G 9505
D. GOUX - E. MAURIN
Persistance des hiérarchies sectorielles de salaires: un réexamen sur données françaises
G 9505
Bis
D. GOUX - E. MAURIN
Persistence of inter-industry wages differentials:
a reexamination on matched worker-firm panel
data
G 9506
S. JACOBZONE
Les liens entre RMI et chômage, une mise en
perspective
NON PARU - article sorti dans Économie et
Prévision n° 122 (1996) - pages 95 à 113
G 9507
G. CETTE - S. MAHFOUZ
Le partage primaire du revenu
Constat descriptif sur longue période
G 9601
Banque de France - CEPREMAP - Direction de
la Prévision - Érasme - INSEE - OFCE
Structures et propriétés de cinq modèles macroéconomiques français
G 9602
Rapport d’activité de la DESE de l’année 1995
G 9603
J. BOURDIEU - A. DRAZNIEKS
L’octroi de crédit aux PME : une analyse à partir
d’informations bancaires
G 9604
A. TOPIOL-BENSAÏD
Les implantations japonaises en France
G 9605
P. GENIER - S. JACOBZONE
Comportements de prévention, consommation
d’alcool et tabagie : peut-on parler d’une gestion
globale du capital santé ?
Une modélisation microéconométrique empirique
G 9606
C. DOZ - F. LENGLART
Factor analysis and unobserved component
models: an application to the study of French
business surveys
G 9607
N. GREENAN - D. GUELLEC
La théorie coopérative de la firme
H. ERKEL-ROUSSE
Le commerce extérieur et l'environnement international dans le modèle AMADEUS
(réestimation 1992)
G 9315
G 9316
D. BLANCHET - C. BROUSSE
Deux études sur l'âge de la retraite
G 9102
J.L. BRILLET
Le modèle AMADEUS - Deuxième partie Propriétés variantielles
G 9207
A. MAGNIER et J. TOUJAS-BERNATE
Technology and trade: empirical evidences for
the major five industrialized countries
G 9317
D. BLANCHET
Répartition du travail dans une population hétérogène : deux notes
G 9103
D. GUELLEC et P. RALLE
Endogenous growth and product innovation
G 9208
G 9318
G 9104
H. ROUSSE
Le modèle AMADEUS - Troisième partie - Le
commerce extérieur et l'environnement
international
B. CREPON, E. DUGUET, D. ENCAOUA et
P. MOHNEN
Cooperative, non cooperative R & D and optimal
patent life
D. EYSSARTIER - N. PONTY
AMADEUS - an annual macro-economic model
for the medium and long term
G 9319
G 9209
B. CREPON et E. DUGUET
Research and development, competition and
innovation: an application of pseudo maximum
likelihood methods to Poisson models with
heterogeneity
G. CETTE - Ph. CUNÉO - D. EYSSARTIER J. GAUTIÉ
Les effets sur l'emploi d'un abaissement du coût
du travail des jeunes
G 9401
D. BLANCHET
Les structures par âge importent-elles ?
J. TOUJAS-BERNATE
Commerce international et concurrence imparfaite : développements récents et implications
pour la politique commerciale
G 9402
J. GAUTIÉ
Le chômage des jeunes en France : problème de
formation ou phénomène de file d'attente ?
Quelques éléments du débat
G 9106
B. CREPON
Innovation, taille et concentration : causalités et
dynamiques
G 9107
B. AMABLE et D. GUELLEC
Un panorama des théories de la croissance
endogène
G 9302
Ch. CASES
Durées de chômage et comportements d'offre de
travail : une revue de la littérature
G 9403
P. QUIRION
Les déchets en France : éléments statistiques et
économiques
G 9108
M. GLAUDE et M. MOUTARDIER
Une évaluation du coût direct de l'enfant de 1979
à 1989
G 9303
H. ERKEL-ROUSSE
Union économique et monétaire : le débat
économique
G 9404
D. LADIRAY - M. GRUN-REHOMME
Lissage par moyennes mobiles - Le problème
des extrémités de série
G 9109
P. RALLE et alii
France - Allemagne : performances économiques comparées
G 9304
G 9405
V. MAILLARD
Théorie et pratique de la correction des effets de
jours ouvrables
G 9110
J.L. BRILLET
Micro-DMS
N. GREENAN - D. GUELLEC /
G. BROUSSAUDIER - L. MIOTTI
Innovation organisationnelle, dynamisme technologique et performances des entreprises
G 9305
P. JAILLARD
Le traité de Maastricht : présentation juridique et
historique
G 9406
F. ROSENWALD
La décision d'investir
NON PARU
G 9111
A. MAGNIER
Effets accélérateur et multiplicateur en France
depuis 1970 : quelques résultats empiriques
G 9112
B. CREPON et G. DUREAU
Investissement en recherche-développement :
analyse de causalités dans un modèle d'accélérateur généralisé
G 9113
G 9201
G 9202
J.L. BRILLET, H. ERKEL-ROUSSE, J. TOUJASBERNATE
"France-Allemagne Couplées" - Deux économies
vues par une maquette macro-économétrique
W.J. ADAMS, B. CREPON, D. ENCAOUA
Choix technologiques et stratégies de dissuasion
d'entrée
J. OLIVEIRA-MARTINS,
J. TOUJAS-BERNATE
G 9306
J.L. BRILLET
Micro-DMS : présentation et propriétés
G 9407
S. JACOBZONE
Les apports de l'économie industrielle pour définir la stratégie économique de l'hôpital public
G 9408
L. BLOCH, J. BOURDIEU,
B. COLIN-SEDILLOT, G. LONGUEVILLE
Du défaut de paiement au dépôt de bilan : les
banquiers face aux PME en difficulté
G 9307
J.L. BRILLET
Micro-DMS - variantes : les tableaux
G 9308
S. JACOBZONE
Les grands réseaux publics français dans une
perspective européenne
G 9409
L. BLOCH - B. CŒURE
Profitabilité de l'investissement productif et
transmission des chocs financiers
D. EYSSARTIER, P. MAIRE
Impacts macro-économiques de mesures d'aide
au logement - quelques éléments d'évaluation
G 9410
F. ROSENWALD
Suivi conjoncturel de l'investissement
G 9411
C. DEFEUILLEY - Ph. QUIRION
Les déchets d'emballages ménagers : une
analyse économique des politiques française et
allemande
G 9309
G 9310
G 9311
J. BOURDIEU - B. COLIN-SEDILLOT
Les théories sur la structure optimale du capital :
quelques points de repère
J. BOURDIEU - B. COLIN-SEDILLOT
Les décisions de financement des entreprises
B. DORMONT - M. PAUCHET
L'évaluation de l'élasticité emploi-salaire dépendelle des structures de qualification ?
G 9313
N. GREENAN et D. GUELLEC
Coordination within the firm and endogenous
growth
G 9301
G 9413
L. BLOCH - B. CŒURÉ
Q de Tobin marginal et transmission des chocs
financiers
G 9206
H. ROUSSE
Effets de demande et d'offre dans les résultats
du commerce extérieur manufacturé de la France
au cours des deux dernières décennies
J. BOURDIEU - B. CŒURÉ B. COLIN-SEDILLOT
Investissement, incertitude et irréversibilité
Quelques développements récents de la théorie
de l'investissement
G 9312
Équipe AMADEUS
Le modèle AMADEUS - Première partie Présentation générale
G 9105
G 9412
iii
iv
G 9608
N. GREENAN - D. GUELLEC
Technological innovation and employment
reallocation
G 9714
F. LEQUILLER
Does the French Consumer Price Index Overstate Inflation?
G 9808
A. MOUROUGANE
Can a Conservative Governor Conduct an Accomodative Monetary Policy?
G 9913
Division « Redistribution et Politiques Sociales »
Le modèle de microsimulation dynamique
DESTINIE
G 9609
Ph. COUR - F. RUPPRECHT
L’intégration asymétrique au sein du continent
américain : un essai de modélisation
G 9715
G 9809
X. BONNET - E. DUBOIS - L. FAUVET
Asymétrie des inflations relatives et menus costs
: tests sur l’inflation française
G 9914
E. DUGUET
Macro-commandes SAS pour l’économétrie des
panels et des variables qualitatives
G 9610
S. DUCHENE - G. FORGEOT - A. JACQUOT
Analyse des évolutions récentes de la productivité apparente du travail
X. BONNET
Peut-on mettre en évidence les rigidités à la
baisse des salaires nominaux ?
Une étude sur quelques grands pays de l’OCDE
G 9810
N. IUNG - F. RUPPRECHT
Productivité de la recherche et rendements
d’échelle dans le secteur pharmaceutique
français
E. DUGUET - N. IUNG
Sales and Advertising with Spillovers at the firm
level: Estimation of a Dynamic Structural Model
on Panel Data
G 9915
G 9716
R. DUHAUTOIS
Évolution des flux d’emplois en France entre
1990 et 1996 : une étude empirique à partir du
fichier des bénéfices réels normaux (BRN)
G 9811
E. DUGUET - I. KABLA
Appropriation strategy and the motivations to use
the patent system in France - An econometric
analysis at the firm level
J.P. BERTHIER
Congestion urbaine : un modèle de trafic de
pointe à courbe débit-vitesse et demande
élastique
G 9916
G 9717
J.Y. FOURNIER
Extraction du cycle des affaires : la méthode de
Baxter et King
G 9917
G 9812
C. PRIGENT
La part des salaires dans la valeur ajoutée : une
approche macroéconomique
B. CRÉPON - R. DESPLATZ - J. MAIRESSE
Estimating price cost margins, scale economies
and workers’ bargaining power at the firm level
G 9918
G 9813
A.Th. AERTS
L’évolution de la part des salaires dans la valeur
ajoutée en France reflète-t-elle les évolutions
individuelles sur la période 1979-1994 ?
Ch. GIANELLA - Ph. LAGARDE
Productivity of hours in the aggregate production
function: an evaluation on a panel of French
firms from the manufacturing sector
G 9919
G 9814
B. SALANIÉ
Guide pratique des séries non-stationnaires
S. AUDRIC - P. GIVORD - C. PROST
Évolution de l’emploi et des coûts par qualification entre 1982 et 1996
G 9901
S. DUCHÊNE - A. JACQUOT
Une croissance plus riche en emplois depuis le
début de la décennie ? Une analyse en comparaison internationale
G 2000/01
R. MAHIEU
Les déterminants des dépenses de santé : une
approche macroéconomique
G 2000/02
G 9902
Ch. COLIN
Modélisation des carrières dans Destinie
G 9903
Ch. COLIN
Évolution de la dispersion des salaires : un essai
de prospective par microsimulation
C. ALLARD-PRIGENT - H. GUILMEAU A. QUINET
The real exchange rate as the relative price of
nontrables in terms of tradables: theoretical
investigation and empirical study on French data
G 2000/03
G 9904
B. CREPON - N. IUNG
Innovation, emploi et performances
J.-Y. FOURNIER
L’approximation du filtre passe-bande proposée
par Christiano et Fitzgerald
G 9905
B. CREPON - Ch. GIANELLA
Wages inequalities in France 1969-1992
An application of quantile regression techniques
G 2000/04
Bilan des activités de la DESE - 1999
G 2000/05
B. CREPON - F. ROSENWALD
Investissement et contraintes de financement : le
poids du cycle
Une estimation sur données françaises
G 2000/06
A. FLIPO
Les comportements matrimoniaux de fait
G 9611
G 9612
G 9613
G 9614
G 9701
G 9702
G 9703
G 9704
G 9705
G 9706
X. BONNET - S. MAHFOUZ
The influence of different specifications of
wages-prices spirals on the measure of the
NAIRU: the case of France
PH. COUR - E. DUBOIS, S. MAHFOUZ,
J. PISANI-FERRY
The cost of fiscal retrenchment revisited: how
strong is the evidence?
G 9718
L.P. PELÉ - P. RALLE
Âge de la retraite : les aspects incitatifs du régime général
G 9719
ZHANG Yingxiang - SONG Xueqing
Lexique macroéconomique Français-Chinois
ZHANG Yingxiang - SONG Xueqing
Lexique macroéconomique français-chinois,
chinois-français
G 9720
J.L. SCHNEIDER
La taxe professionnelle : éléments de cadrage
économique
M. HOUDEBINE - J.L. SCHNEIDER
Mesurer l’influence de la fiscalité sur la localisation des entreprises
G 9721
J.L. SCHNEIDER
Transition et stabilité politique d’un système
redistributif
A. MOUROUGANE
Crédibilité, indépendance et politique monétaire
Une revue de la littérature
G 9722
P. AUGERAUD - L. BRIOT
Les données comptables d’entreprises
Le système intermédiaire d’entreprises
Passage des données individuelles aux données
sectorielles
A. JACQUOT
Les flexions des taux d’activité sont-elles seulement conjoncturelles ?
D. GOUX - E. MAURIN
Train or Pay: Does it Reduce Inequalities to Encourage Firms to Train their Workers?
P. GENIER
Deux contributions sur dépendance et équité
G 9723
E. DUGUET - N. IUNG
R & D Investment, Patent Life and Patent Value
An Econometric Analysis at the Firm Level
P. AUGERAUD - J.E. CHAPRON
Using Business Accounts for Compiling National
Accounts: the French Experience
G 9724
P. AUGERAUD
Les comptes d’entreprise par activités - Le passage aux comptes - De la comptabilité
d’entreprise à la comptabilité nationale - A
paraître
M. HOUDEBINE - A. TOPIOL-BENSAÏD
Les entreprises internationales en France : une
analyse à partir de données individuelles
G 9707
M. HOUDEBINE
Polarisation des activités et spécialisation des
départements en France
G 9708
E. DUGUET - N. GREENAN
Le biais technologique : une analyse sur données individuelles
G 9709
J.L. BRILLET
Analyzing a small French ECM Model
G 9710
J.L. BRILLET
Formalizing the transition process: scenarios for
capital accumulation
G 9711
G. FORGEOT - J. GAUTIÉ
Insertion professionnelle des jeunes et processus de déclassement
G 9712
E. DUBOIS
High Real Interest Rates: the Consequence of a
Saving Investment Disequilibrium or of an insufficient Credibility of Monetary Authorities?
G 9713
G 9801
H. MICHAUDON - C. PRIGENT
Présentation du modèle AMADEUS
G 9802
J. ACCARDO
Une étude de comptabilité générationnelle
pour la France en 1996
G 9803
X. BONNET - S. DUCHÊNE
Apports et limites de la modélisation
« Real Business Cycles »
G 9804
G 9805
G 9806
Bilan des activités de la Direction des Études
et Synthèses Économiques - 1996
G 9807
C. BARLET - C. DUGUET D. ENCAOUA - J. PRADEL
The Commercial Success of Innovations
An econometric analysis at the firm level in
French manufacturing
P. CAHUC - Ch. GIANELLA D. GOUX - A. ZILBERBERG
Equalizing Wage Differences and Bargaining
Power - Evidence form a Panel of French Firms
J. ACCARDO - M. JLASSI
La productivité globale des facteurs entre 1975
et 1996
Bilan des activités de la Direction des Études et
Synthèses Économiques - 1997
G 9906
C. BONNET - R. MAHIEU
Microsimulation techniques applied to intergenerational transfers - Pensions in a dynamic
framework: the case of France
G 9907
F. ROSENWALD
L’impact des contraintes financières dans la décision d’investissement
G 2000/07
R. MAHIEU - B. SÉDILLOT
Microsimulations of the retirement decision: a
supply side approach
G 9908
Bilan des activités de la DESE - 1998
G 2000/08
G 9909
J.P. ZOYEM
Contrat d’insertion et sortie du RMI
Évaluation des effets d’une politique sociale
C. AUDENIS - C. PROST
Déficit conjoncturel : une prise en compte des
conjonctures passées
G 2000/09
G 9910
Ch. COLIN - Fl. LEGROS - R. MAHIEU
Bilans contributifs comparés des régimes de
retraite du secteur privé et de la fonction
publique
R. MAHIEU - B. SÉDILLOT
Équivalent patrimonial de la rente et souscription
de retraite complémentaire
G 2000/10
R. DUHAUTOIS
Ralentissement de l’investissement : petites ou
grandes entreprises ? industrie ou tertiaire ?
G 9911
G. LAROQUE - B. SALANIÉ
Une décomposition du non-emploi en France
G 2000/11
G 9912
B. SALANIÉ
Une maquette analytique de long terme du
marché du travail
G. LAROQUE - B. SALANIÉ
Temps partiel féminin et incitations financières à
l’emploi
G2000/12
G 9912
Bis
Ch. GIANELLA
Une estimation de l’élasticité de l’emploi peu
qualifié à son coût
Ch. GIANELLA
Local unemployment and wages
G2000/13
B. CREPON - Th. HECKEL
- Informatisation en France : une évaluation à
partir de données individuelles
v
- Computerization in France: an evaluation based
on individual company data
G2001/01
G2001/02
G2001/03
G2001/04
G2001/05
F. LEQUILLER
- La nouvelle économie et la mesure
de la croissance du PIB
- The new economy and the measure
ment of GDP growth
S. AUDRIC
La reprise de la croissance de l’emploi profite-telle aussi aux non-diplômés ?
A. BEAUDU - Th. HECKEL
Le canal du crédit fonctionne-t-il en Europe ?
Une étude de l’hétérogénéité des comportements d’investissement à partir de données
de bilan agrégées
C. AUDENIS - P. BISCOURP N. FOURCADE - O. LOISEL
Testing the augmented Solow growth model: An
empirical reassessment using panel data
R. MAHIEU - B. SÉDILLOT
Départ à la retraite, irréversibilité et incertitude
G2001/07
Bilan des activités de la DESE - 2000
G2001/08
J. Ph. GAUDEMET
Les dispositifs d’acquisition à titre facultatif
d’annuités viagères de retraite
G2001/10
G2001/11
G2001/12
G2001/13
G2001/14
G2002/01
F. MAGNIEN - J.-L. TAVERNIER - D. THESMAR
Les statistiques internationales de PIB par
habitant en standard de pouvoir d’achat : une
analyse des résultats
G2002/02
Bilan des activités de la DESE - 2001
G2002/03
B. SÉDILLOT - E. WALRAET
La cessation d’activité au sein des couples : y at-il interdépendance des choix ?
G2002/04
G. BRILHAULT
- Rétropolation des séries de FBCF et calcul du
capital fixe en SEC-95 dans les comptes
nationaux français
- Retropolation of the investment series (GFCF)
and estimation of fixed capital stocks on the
ESA-95 basis for the French balance sheets
I. BRAUN-LEMAIRE
Évolution et répartition du surplus de productivité
G2001/06
G2001/09
vi
B. CRÉPON - Ch. GIANELLA
Fiscalité, coût d’usage du capital et demande de
facteurs : une analyse sur données individuelles
B. CRÉPON - R. DESPLATZ
Évaluation
des
effets
des
dispositifs
d’allégements
de charges sociales sur les bas salaires
G2002/05
G2002/06
G2002/07
P. BISCOURP - Ch. GIANELLA
Substitution and complementarity between
capital, skilled and less skilled workers: an
analysis at the firm level in the French
manufacturing industry
I. ROBERT-BOBEE
Modelling demographic behaviours in the French
microsimulation model Destinie: An analysis of
future change in completed fertility
G2001/15
J.-P. ZOYEM
Diagnostic sur la pauvreté et calendrier de
revenus : le cas du “Panel européen des
ménages »
G2001/16
J.-Y. FOURNIER - P. GIVORD
La réduction des taux d’activité aux âges
extrêmes, une spécificité française ?
G2001/17
C. AUDENIS - P. BISCOURP - N. RIEDINGER
Existe-t-il une asymétrie dans la transmission du
prix du brut aux prix des carburants ?
C. AUDENIS - J. DEROYON - N. FOURCADE
L’impact des nouvelles technologies de
l’information et de la communication sur
l’économie française - un bouclage macroéconomique
J. BARDAJI - B. SÉDILLOT - E. WALRAET
Évaluation de trois réformes du Régime Général
d’assurance vieillesse à l’aide du modèle de
microsimulation DESTINIE
G2002/08
J.-P. BERTHIER
Réflexions sur les différentes notions de volume
dans les comptes nationaux : comptes aux prix
d’une année fixe ou aux prix de l’année
précédente, séries chaînées
G2002/09
F. HILD
Les soldes d’opinion résument-ils au mieux les
réponses des entreprises aux enquêtes de
conjoncture ?
G2002/10
I. ROBERT-BOBÉE
Les comportements démographiques dans le
modèle de microsimulation Destinie - Une
comparaison des estimations issues des
enquêtes Jeunes et Carrières 1997 et Histoire
Familiale 1999
J.-Y. FOURNIER
Comparaison des salaires des secteurs public et
privé
J.-P. BERTHIER - C. JAULENT
R. CONVENEVOLE - S. PISANI
Une méthodologie de comparaison entre
consommations intermédiaires de source fiscale
et de comptabilité nationale
P. BISCOURP - B. CRÉPON - T. HECKEL - N.
RIEDINGER
How do firms respond to cheaper computers?
Microeconometric evidence for France based on
a production function approach
G2002/11
J.-P. ZOYEM
La dynamique des bas revenus : une analyse
des entrées-sorties de pauvreté
G2002/16
F. MAUREL - S. GREGOIR
Les indices de compétitivité des pays : interprétation et limites
G2004/06
M. DUÉE
L’impact du chômage des parents sur le devenir
scolaire des enfants
G2003/01
N. RIEDINGER - E.HAUVY
Le coût de dépollution atmosphérique pour les
entreprises françaises : Une estimation à partir
de données individuelles
G2004/07
P. AUBERT - E. CAROLI - M. ROGER
New Technologies, Workplace Organisation and
the Age Structure of the Workforce: Firm-Level
Evidence
G2003/02
P. BISCOURP et F. KRAMARZ
Création d’emplois, destruction d’emplois et
internationalisation des entreprises industrielles
françaises : une analyse sur la période 19861992
G2004/08
E. DUGUET - C. LELARGE
Les brevets accroissent-ils les incitations privées
à innover ? Un examen microéconométrique
G2004/09
G2003/03
Bilan des activités de la DESE - 2002
S. RASPILLER - P. SILLARD
Affiliating versus Subcontracting:
the Case of Multinationals
G2003/04
P.-O. BEFFY - J. DEROYON N. FOURCADE - S. GREGOIR - N. LAÏB B. MONFORT
Évolutions démographiques et croissance : une
projection macro-économique à l’horizon 2020
G2004/10
J. BOISSINOT - C. L’ANGEVIN - B. MONFORT
Public Debt Sustainability: Some Results on the
French Case
G2004/11
G2003/05
P. AUBERT
La situation des salariés de plus de cinquante
ans dans le secteur privé
S. ANANIAN - P. AUBERT
Travailleurs âgés, nouvelles technologies
et changements organisationnels : un réexamen
à partir de l’enquête « REPONSE »
G2004/12
G2003/06
P. AUBERT - B. CRÉPON
Age, salaire et productivité
La productivité des salariés décline-t-elle en fin
de carrière ?
X. BONNET - H. PONCET
Structures de revenus et propensions différentes
à consommer - Vers une équation de
consommation des ménages plus robuste en
prévision pour la France
G2003/07
H. BARON - P.O. BEFFY - N. FOURCADE - R.
MAHIEU
Le ralentissement de la productivité du travail au
cours des années 1990
G2004/13
C. PICART
Évaluer la
financières
G2003/08
P.-O. BEFFY - B. MONFORT
Patrimoine des ménages, dynamique d’allocation
et comportement de consommation
G2003/09
P. BISCOURP - N. FOURCADE
Peut-on mettre en évidence l’existence de
rigidités à la baisse des salaires à partir de
données individuelles ? Le cas de la France à la
fin des années 90
G2003/10
M. LECLAIR - P. PETIT
Présence syndicale dans les firmes : quel impact
sur les inégalités salariales entre les hommes et
les femmes ?
P.-O. BEFFY - X. BONNET - M. DARRACQPARIES - B. MONFORT
MZE: a small macro-model for the euro area
G2003/11
G2004/01
P. AUBERT - M. LECLAIR
La compétitivité exprimée dans les enquêtes
trimestrielles sur la situation et les perspectives
dans l’industrie
G2002/12
F. HILD
Prévisions d’inflation pour la France
G2002/13
M. LECLAIR
Réduction du temps de travail et tensions sur les
facteurs de production
G2004/02
M. DUÉE - C. REBILLARD
La dépendance des personnes âgées : une
projection à long terme
E. WALRAET - A. VINCENT
- Analyse de la redistribution intragénérationnelle
dans le système de retraite des salariés du privé
- Une approche par microsimulation
- Intragenerational distributional analysis in the
french private sector pension scheme - A
microsimulation approach
G2004/03
S. RASPILLER - N. RIEDINGER
Régulation environnementale et
localisation des groupes français
G2004/04
A. NABOULET - S. RASPILLER
Les déterminants de la décision d’investir : une
approche par les perceptions subjectives des
firmes
P. CHONE - D. LE BLANC - I. ROBERT-BOBEE
Offre de travail féminine et garde des jeunes
enfants
G2004/05
N. RAGACHE
La déclaration des enfants par les couples non
mariés est-elle fiscalement optimale ?
G2002/14
G2002/15
choix
de
rentabilité
des
sociétés
non
G2004/14
J. BARDAJI - B. SÉDILLOT - E. WALRAET
Les retraites du secteur public : projections à
l’horizon 2040 à l’aide du modèle de
microsimulation DESTINIE
G2005/01
S. BUFFETEAU - P. GODEFROY
Conditions de départ en retraite selon l’âge de fin
d’études : analyse prospective pour les
générations 1945 à1974
G2005/02
C. AFSA - S. BUFFETEAU
L’évolution de l’activité féminine en France :
une approche par pseudo-panel
G2005/03
P. AUBERT - P. SILLARD
Délocalisations et réductions d’effectifs
dans l’industrie française
G2005/04
M. LECLAIR - S. ROUX
Mesure et utilisation des emplois instables
dans les entreprises
G2005/05
C. L’ANGEVIN - S. SERRAVALLE
Performances à l’exportation de la France
et de l’Allemagne - Une analyse par secteur et
destination géographique
G2005/06
Bilan des activités de la Direction des Études et
Synthèses Économiques - 2004
G2005/07
S. RASPILLER
La concurrence fiscale : principaux enseignements de l’analyse économique
G2005/08
C. L’ANGEVIN - N. LAÏB
Éducation et croissance en France et dans un
panel de 21 pays de l’OCDE
G2005/09
N. FERRARI
Prévoir l’investissement des entreprises
Un indicateur des révisions dans l’enquête de
conjoncture sur les investissements dans
l’industrie.
vii
viii
G2009/09
D. BLANCHET - F. LE GALLO
Les projections démographiques : principaux
mécanismes et retour sur l’expérience française
G. LALANNE - E. POULIQUEN - O. SIMON
Prix du pétrole et croissance potentielle à long
terme
G2009/10
D. BLANCHET - F. TOUTLEMONDE
Évolutions démographiques et déformation du
cycle de vie active : quelles relations ?
D. BLANCHET - J. LE CACHEUX - V. MARCUS
Adjusted net savings and other approaches to
sustainability: some theoretical background
G2009/11
V. BELLAMY - G. CONSALES - M. FESSEAU S. LE LAIDIER - É. RAYNAUD
Une décomposition du compte des ménages de
la comptabilité nationale par catégorie de
ménage en 2003
G2009/12
J. BARDAJI - F. TALLET
Detecting Economic Regimes in France: a
Qualitative Markov-Switching Indicator Using
Mixed Frequency Data
G2009/13
R.
AEBERHARDT
D.
FOUGÈRE
R. RATHELOT
Discrimination à l’embauche : comment exploiter
les procédures de testing ?
G2009/14
Y. BARBESOL - P. GIVORD - S. QUANTIN
Partage de la valeur ajoutée, approche par
données microéconomiques
G2009/15
I. BUONO - G. LALANNE
The Effect of the Uruguay round on the Intensive
and Extensive Margins of Trade
G2010/01
C. MINODIER
Avantages comparés des séries des premières
valeurs publiées et des séries des valeurs
révisées - Un exercice de prévision en temps réel
de la croissance trimestrielle du PIB en France
G2010/02
V. ALBOUY - L. DAVEZIES - T. DEBRAND
Health Expenditure Models: a Comparison of
Five Specifications using Panel Data
G2010/03
C. KLEIN - O. SIMON
Le modèle MÉSANGE réestimé en base 2000
Tome 1 – Version avec volumes à prix constants
G2010/04
M.-É. CLERC - É. COUDIN
L’IPC, miroir de l’évolution du coût de la vie en
France ? Ce qu’apporte l’analyse des courbes
d’Engel
G2010/05
N. CECI-RENAUD - P.-A. CHEVALIER
Les seuils de 10, 20 et 50 salariés : impact sur la
taille des entreprises françaises
G2010/06
R. AEBERHARDT - J. POUGET
National Origin Differences in Wages and
Hierarchical Positions - Evidence on French FullTime Male Workers from a matched EmployerEmployee Dataset
G2010/07
S. BLASCO - P. GIVORD
Les trajectoires professionnelles en début de vie
active : quel impact des contrats temporaires ?
G2010/08
P. GIVORD
Méthodes économétriques pour l’évaluation de
politiques publiques
G2010/09
P.-Y. CABANNES - V. LAPÈGUE E. POULIQUEN - M. BEFFY - M. GAINI
Quelle croissance de moyen terme après la
crise ?
G2010/10
I. BUONO - G. LALANNE
La réaction des entreprises françaises
à la baisse des tarifs douaniers étrangers
G2005/10
P.-O. BEFFY - C. L’ANGEVIN
Chômage et boucle prix-salaires :
apport d’un modèle « qualifiés/peu qualifiés »
G2006/11
C. LELARGE
Les entreprises (industrielles) françaises sontelles à la frontière technologique ?
G2005/11
B. HEITZ
A two-states Markov-switching model of inflation
in France and the USA: credible target VS
inflation spiral
G2006/12
O. BIAU - N. FERRARI
Théorie de l’opinion
Faut-il pondérer les réponses individuelles ?
G2006/13
G2005/12
O. BIAU - H. ERKEL-ROUSSE - N. FERRARI
Réponses individuelles aux enquêtes de
conjoncture et prévision macroéconomiques :
Exemple de la prévision de la production
manufacturière
A. KOUBI - S. ROUX
Une réinterprétation de la relation entre
productivité et inégalités salariales dans les
entreprises
G2008/06
R. RATHELOT - P. SILLARD
The impact of local taxes on plants location
decision
M. BARLET - D. BLANCHET - L. CRUSSON
Internationalisation et flux d’emplois : que dit une
approche comptable ?
G2008/07
C. LELARGE - D. SRAER - D. THESMAR
Entrepreneurship and Credit Constraints Evidence from a French Loan Guarantee
Program
G2008/08
X. BOUTIN - L. JANIN
Are Prices Really Affected by Mergers?
G2008/09
M. BARLET - A. BRIANT - L. CRUSSON
Concentration géographique dans l’industrie
manufacturière et dans les services en France :
une approche par un indicateur en continu
G2005/13
G2005/14
P. AUBERT - D. BLANCHET - D. BLAU
The labour market after age 50: some elements
of a Franco-American comparison
D. BLANCHET - T. DEBRAND P. DOURGNON - P. POLLET
L’enquête SHARE : présentation et premiers
résultats de l’édition française
G2005/15
M. DUÉE
La modélisation des comportements démographiques dans le modèle de microsimulation
DESTINIE
G2005/16
G2006/14
L. GONZALEZ - C. PICART
Diversification, recentrage et poids des activités
de support dans les groupes (1993-2000)
G2007/01
D. SRAER
Allègements de cotisations
dynamique salariale
et
G2007/02
V. ALBOUY - L. LEQUIEN
Les rendements non monétaires de l’éducation :
le cas de la santé
H. RAOUI - S. ROUX
Étude de simulation sur la participation versée
aux salariés par les entreprises
G2007/03
D. BLANCHET - T. DEBRAND
Aspiration à la retraite, santé et satisfaction au
travail : une comparaison européenne
G2006/01
C. BONNET - S. BUFFETEAU - P. GODEFROY
Disparités de retraite de droit direct entre
hommes et femmes : quelles évolutions ?
G2007/04
G2006/02
C. PICART
Les gazelles en France
G2007/05
G2006/03
P. AUBERT - B. CRÉPON -P. ZAMORA
Le rendement apparent de la formation continue
dans les entreprises : effets sur la productivité et
les salaires
G2006/04
G2006/05
G2006/06
G2006/07
G2006/08
G2006/09
G2006/10
G2007/06
G2008/04
G2008/05
G2006/15
patronales
entreprises : estimation sur données individuelles
françaises
G2008/10
M. BEFFY - É. COUDIN - R. RATHELOT
Who is confronted to insecure labor market
histories? Some evidence based on the French
labor market transition
M. BARLET - L. CRUSSON
Quel impact des variations du prix du pétrole sur
la croissance française ?
G2008/11
M. ROGER - E. WALRAET
Social Security and Well-Being of the Elderly: the
Case of France
C. PICART
Flux d’emploi et de main-d’œuvre en France : un
réexamen
G2008/12
C. AFSA
Analyser les composantes du bien-être et de son
évolution
Une
approche
empirique
sur
données
individuelles
V. ALBOUY - C. TAVAN
Massification
et
démocratisation
l’enseignement supérieur en France
de
G2008/13
T. LE BARBANCHON
The Changing response to oil price shocks in
France: a DSGE type approach
M. BARLET - D. BLANCHET T. LE BARBANCHON
Microsimuler le marché du travail : un prototype
G2009/01
P.-A. PIONNIER
Le partage de la valeur ajoutée en France,
1949-2007
J.-F. OUVRARD - R. RATHELOT
Demographic change and unemployment:
what do macroeconometric models predict?
G2007/07
D. BLANCHET - J.-F. OUVRARD
Indicateurs d’engagements implicites des
systèmes de retraite : chiffrages, propriétés
analytiques et réactions à des chocs
démographiques types
G2007/08
T. CHANEY - D. SRAER - D. THESMAR
Collateral Value and Corporate Investment
Evidence from the French Real Estate Market
G2009/02
G2007/09
Laurent CLAVEL - Christelle MINODIER
A Monthly Indicator of the French Business
Climate
G. BIAU - O. BIAU - L. ROUVIERE
Nonparametric Forecasting of the Manufacturing
Output Growth with Firm-level Survey Data
J. BOISSINOT
Consumption over the Life Cycle: Facts for
France
G2009/03
G2007/10
H. ERKEL-ROUSSE - C. MINODIER
Do Business Tendency Surveys in Industry and
Services Help in Forecasting GDP Growth?
A Real-Time Analysis on French Data
C. AFSA - P. GIVORD
Le rôle des conditions de travail dans les
absences pour maladie
C. AFSA
Interpréter les variables de
l’exemple de la durée du travail
G2007/11
G2009/04
P. GIVORD - L. WILNER
Les contrats temporaires : trappe ou marchepied
vers l’emploi stable ?
P. SILLARD - C. L’ANGEVIN - S. SERRAVALLE
Performances comparées à l’exportation de la
France et de ses principaux partenaires
Une analyse structurelle sur 12 ans
X. BOUTIN - S. QUANTIN
Une méthodologie d’évaluation comptable du
coût du capital des entreprises françaises : 19842002
C. AFSA
L’estimation d’un coût implicite de la pénibilité du
travail chez les travailleurs âgés
satisfaction :
R. RATHELOT - P. SILLARD
Zones Franches Urbaines : quels effets sur
l’emploi
salarié
et
les
créations
d’établissements ?
G2007/12
V. ALBOUY - B. CRÉPON
Aléa moral en santé : une évaluation dans le
cadre du modèle causal de Rubin
G2008/01
C. PICART
Les PME françaises :
dynamiques
G2008/02
G2008/03
rentables
mais
G2009/05
G2009/06
peu
P. BISCOURP - X. BOUTIN - T. VERGÉ
The Effects of Retail Regulations on Prices
Evidence form the Loi Galland
Y. BARBESOL - A. BRIANT
Économies d’agglomération et productivité des
G. LALANNE - P.-A. PIONNIER - O. SIMON
Le partage des fruits de la croissance de 1950 à
2008 : une approche par les comptes de surplus
L. DAVEZIES - X. D’HAULTFOEUILLE
Faut-il pondérer ?… Ou l’éternelle question de
l’économètre confronté à des données d’enquête
G2009/07
S. QUANTIN - S. RASPILLER - S. SERRAVALLE
Commerce intragroupe, fiscalité et prix de
transferts : une analyse sur données françaises
G2009/08
M. CLERC - V. MARCUS
Élasticités-prix des consommations énergétiques
des ménages
ix
G2010/11
G2010/12
G2010/13
G2010/14
G2010/15
G2010/16
G2010/17
R. RATHELOT - P. SILLARD
L’apport des méthodes à noyaux pour mesurer la
concentration géographique - Application à la
concentration des immigrés en France de 1968 à
1999
M. BARATON - M. BEFFY - D. FOUGÈRE
Une évaluation de l’effet de la réforme de 2003
sur les départs en retraite - Le cas des
enseignants du second degré public
D. BLANCHET - S. BUFFETEAU - E. CRENNER
S. LE MINEZ
Le modèle de microsimulation Destinie 2 :
principales caractéristiques et premiers résultats
x
prises sur la base des contrôles fiscaux et son
insertion dans les comptes nationaux
G2011/10
G2011/11
G2011/12
D. BLANCHET - E. CRENNER
Le bloc retraites du modèle Destinie 2 :
guide de l’utilisateur
G2011/13
M. BARLET - L. CRUSSON - S. DUPUCH F. PUECH
Des services échangés aux services échangeables : une application sur données françaises
G2011/14
M. BEFFY - T. KAMIONKA
Public-private wage gaps: is civil-servant human
capital sector-specific?
P.-Y. CABANNES - H. ERKEL-ROUSSE G. LALANNE - O. MONSO - E. POULIQUEN
Le modèle Mésange réestimé en base 2000
Tome 2 - Version avec volumes à prix chaînés
G2011/15
G2011/16
A. SCHREIBER - A. VICARD
La tertiarisation de l’économie française et le
ralentissement de la productivité entre 1978 et
2008
M.-É. CLERC - O. MONSO - E. POULIQUEN
Les inégalités entre générations depuis le babyboom
C. MARBOT - D. ROY
Évaluation de la transformation de la réduction
d'impôt en crédit d'impôt pour l'emploi de salariés
à domicile en 2007
P. GIVORD - R. RATHELOT - P. SILLARD
Place-based tax exemptions and displacement
effects: An evaluation of the Zones Franches
Urbaines program
G2012/10
C. MARBOT - D. ROY
Projections du coût de l’APA et des
caractéristiques de ses bénéficiaires à l’horizon
2040 à l’aide du modèle Destinie
G2013/14
A. POISSONNIER - D. ROY
Households Satellite Account for France in 2010.
Methodological issues on the assessment of
domestic production
G2012/11
A. MAUROUX
Le crédit d’impôt dédié au développement
durable : une évaluation économétrique
G2013/15
G. CLÉAUD - M. LEMOINE - P.-A. PIONNIER
Which size and evolution of the government
expenditure multiplier in France (1980-2010)?
G2012/12
V. COTTET - S. QUANTIN - V. RÉGNIER
Coût du travail et allègements de charges : une
estimation au niveau établissement de 1996 à
2008
G2014/01
M. BACHELET - A. LEDUC - A. MARINO
Les biographies du modèle Destinie II : rebasage
et projection
G2014/02
G2012/13
X. D’HAULTFOEUILLE - P. FÉVRIER L. WILNER
Demand Estimation in the Presence of Revenue
Management
B. GARBINTI
L’achat de la résidence principale et la création
d’entreprises sont-ils favorisés par les donations
et héritages ?
G2014/03
G2012/14
D. BLANCHET - S. LE MINEZ
Joint macro/micro evaluations of accrued-to-date
pension liabilities: an application to French
reforms
N. CECI-RENAUD - P. CHARNOZ - M. GAINI
Évolution de la volatilité des revenus salariaux du
secteur privé en France depuis 1968
G2014/04
P. AUBERT
Modalités d’application des réformes des
retraites et prévisibilité du montant de pension
G2014/05
C. GRISLAIN-LETRÉMY - A. KATOSSKY
The Impact of Hazardous Industrial Facilities on
Housing Prices: A Comparison of Parametric and
Semiparametric Hedonic Price Models
X. D’HAULTFOEUILLE - P. GIVORD X. BOUTIN
The Environmental Effect of Green Taxation: the
Case of the French “Bonus/Malus”
G2013/01F1301
M. BARLET - M. CLERC - M. GARNEO V. LAPÈGUE - V. MARCUS
La nouvelle version du modèle MZE, modèle
macroéconométrique pour la zone euro
T. DEROYON - A. MONTAUT - P-A PIONNIER
Utilisation rétrospective de l’enquête Emploi à
une fréquence mensuelle : apport d’une
modélisation espace-état
G2013/02F1302
C. TREVIEN
Habiter en HLM : quel avantage monétaire et
quel impact sur les conditions de logement ?
G2013/03
A. POISSONNIER
Temporal disaggregation of stock variables - The
Chow-Lin method extended to dynamic models
G2013/04
P. GIVORD - C. MARBOT
Does the cost of child care affect female labor
market participation? An evaluation of a French
reform of childcare subsidies
R. AEBERHARDT - I. BUONO - H. FADINGER
Learning, Incomplete Contracts and Export
Dynamics: theory and Evidence form French
Firms
G2010/18
R. AEBERHARDT - L. DAVEZIES
Conditional Logit with one Binary Covariate: Link
between the Static and Dynamic Cases
G2011/17
G2011/01
T. LE BARBANCHON - B. OURLIAC - O. SIMON
Les marchés du travail français et américain face
aux chocs conjoncturels des années 1986 à
2007 : une modélisation DSGE
C. KERDRAIN - V. LAPÈGUE
Restrictive Fiscal Policies in Europe:
What are the Likely Effects?
G2012/01
C. MARBOT
Une évaluation de la réduction d’impôt pour
l’emploi de salariés à domicile
P. GIVORD - S. QUANTIN - C. TREVIEN
A Long-Term Evaluation of the First Generation
of the French Urban Enterprise Zones
G2013/05
G2011/02
G2012/02
N. CECI-RENAUD - V. COTTET
Politique
salariale
et
performance
entreprises
G. LAME - M. LEQUIEN - P.-A. PIONNIER
Interpretation and limits of sustainability tests in
public finance
G2013/06
C. BELLEGO - V. DORTET-BERNADET
La participation aux pôles de compétitivité :
quelle incidence sur les dépenses de R&D et
l’activité des PME et ETI ?
G2013/07
P-Y. CABANNES - A.MONTAUT P-A. PIONNIER
Évaluer la productivité globale des facteurs en
France : l’apport d’une mesure de la qualité du
capital et du travail
G2011/03
L. DAVEZIES
Modèles à effets fixes, à effets aléatoires,
modèles mixtes ou multi-niveaux : propriétés et
mises en œuvre des modélisations de
l’hétérogénéité dans le cas de données groupées
G2012/03
P. FÉVRIER - L. WILNER
Do
Consumers
Correctly
Expect
Reductions? Testing Dynamic Behavior
des
Price
G2012/04
M. GAINI - A. LEDUC - A. VICARD
School as a shelter? School leaving-age and the
business cycle in France
J.-C. BRICONGNE - J.-M. FOURNIER
V. LAPÈGUE - O. MONSO
De la crise financière à la crise économique
L’impact des perturbations financières de 2007 et
2008 sur la croissance de sept pays
industrialisés
G2012/05
M. GAINI - A. LEDUC - A. VICARD
A scarred generation? French evidence on young
people entering into a tough labour market
G2013/08
R. AEBERHARDT - C. MARBOT
Evolution of Instability on the French Labour
Market During the Last Thirty Years
G2012/06
P. AUBERT - M. BACHELET
Disparités de montant de pension et
redistribution dans le système de retraite français
G2013/09
J-B. BERNARD - G. CLÉAUD
Oil price: the nature of the shocks and the impact
on the French economy
G2011/06
P. CHARNOZ - É. COUDIN - M. GAINI
Wage inequalities in France 1976-2004:
a quantile regression analysis
G2012/07
R. AEBERHARDT - P GIVORD - C. MARBOT
Spillover Effect of the Minimum Wage in France:
An Unconditional Quantile Regression Approach
G2013/10
G. LAME
Was there a « Greenspan Conundrum » in the
Euro area?
G2011/07
M. CLERC - M. GAINI - D. BLANCHET
Recommendations of the Stiglitz-Sen-Fitoussi
Report: A few illustrations
G2012/08
G2013/11
P. CHONÉ - F. EVAIN - L. WILNER - E. YILMAZ
Introducing activity-based payment in the
hospital industry : Evidence from French data
G2011/08
M. BACHELET - M. BEFFY - D. BLANCHET
Projeter l’impact des réformes des retraites sur
l’activité des 55 ans et plus : une comparaison de
trois modèles
A. EIDELMAN - F. LANGUMIER - A. VICARD
Prélèvements obligatoires reposant sur les
ménages : des canaux redistributifs différents en
1990 et 2010
G2012/09
O. BARGAIN - A. VICARD
Le RMI et son successeur le RSA découragentils certains jeunes de travailler ? Une analyse sur
les jeunes autour de 25 ans
G2013/12
C. GRISLAIN-LETRÉMY
Natural Disasters: Exposure and Underinsurance
G2013/13
P.-Y. CABANNES - V. COTTET - Y. DUBOIS C. LELARGE - M. SICSIC
French Firms in the Face of the 2008/2009 Crisis
G2011/04
G2011/05
G2011/09
M. ROGER - M. WASMER
Heterogeneity matters: labour
differentiated by age and skills
productivity
C. LOUVOT-RUNAVOT
L’évaluation de l’activité dissimulée des entre-

IOTA P-4p-Fongicide-09-2014:DT EOLE-Moyen

Téléchagez le journal en format PDF - Municipalité de St

90 - Point Vert

90 - Point Vert

Retrouvez la liste des exposants sur notre catalogue 2014

Are You suprised ?

Télécharger - Jardin Express